M 80 60 40 20 10
Y 80 60 40 20 10 C 80 60 40 20 10 B 80 60 40 20 10
M 80 60 40 20 10
Y 80 60 40 20 10 C 80 60 40 20 10 B 80 60 40 20 10
Ecology of the City of Sofia
Y 80 60 40 20 10 C 80 60 40 20 10 B 80 60 40 20 10
Y 80 60 40 20 10 C 80 60 40 20 10 B 80 60 40 20 10
M 80 60 40 20 10
Y 80 60 40 20 10 C 80 60 40 20 10 B 80 60 40 20 10
M 80 60 40 20 10
Y 80 60 40 20 10 C 80 60 40 20 10 B 80 60 40 20 10
It is expected that in this year
Penev Niemelä Kotze Chipev Editors
research on the biota and to outline some recommendations for improving human comfort in the city. The botanical part consists of nine papers, ranging from inventories (bryophytes, vascular plants) to studies on the soil moisture regime and lead transformation in city parks. In the third part, eleven papers summarize data on several animal groups – nematodes, myriapods, terrestrial gastropods, opilionids, spiders, carabid beetles, ants, aphids, scuttle-flies and birds. The book is intended for urban ecologists, conservationists, environmentalists, urban planning specialists, zoologists, botanists and anyone interested in ecological and environmental issues.
M 80 60 40 20 10
M 80 60 40 20 10
SPECIES AND COMMUNITIES IN AN URBAN ENVIRONMENT Edited by L. Penev, J. Niemelä, D. J. Kotze & N. Chipev
M 80 60 40 20 10
Y 80 60 40 20 10 C 80 60 40 20 10 B 80 60 40 20 10
(2005), about half of the human population will be urban dwellers, and the proportion is increasing. Thus, urban areas – including green areas – are becoming increasingly important for humankind. In order to manage urban areas for the benefit of people and biodiversity, more information is needed about ecological patterns and processes in cities. The present volume is a result of many years of study of the environment and biota of the city of Sofia. The idea of the book came about from the participants of the international GLOBENET project aiming at studying landscape changes in urban environment with both standardized sampling and data methods in various cities across the world using the “urban-torural gradient” concept. The book consists of three parts – Environment, Plants and Animals. The papers in the first part aim to describe the environment of Sofia – climate, water and air – in order to provide a basis for present-day and future
M 80 60 40 20 10
Y 80 60 40 20 10 C 80 60 40 20 10 B 80 60 40 20 10
Ecology of the Sofia City Page i
Contents
ECOLOGY OF THE CITY OF SOFIA Species and Communities in an Urban Environment
i
Ecology of the Sofia City Page ii
ii
L. Penev, J. Niemelä, D. J. Kotze & N. Chipev
This page intentionally left blank
Ecology of the Sofia City Page iii
Contents
iii
Ecology of the City of
Sofia
Species and Communities in an Urban Environment Edited by L. Penev, J. Niemelä, D. J. Kotze & N. Chipev
Sofia-Moscow 2004
Ecology of the Sofia City Page iv
iv
L. Penev, J. Niemelä, D. J. Kotze & N. Chipev
ECOLOGY OF THE CITY OF SOFIA Species and Communities in an Urban Environment Editors: Dr. LYUBOMIR PENEV, Associate Professor Central Laboratory of General Ecology Bulgarian Academy of Sciences 2, Gagarin Str., 1113 Sofia, Bulgaria
[email protected] Dr. JARI NIEMELÄ, Professor Department of Biological and Environmental Sciences Faculty of Biosciences P.O. Box 65, FI-00014 University of Helsinki, Finland
[email protected] Dr. JOHAN KOTZE Department of Biological and Environmental Sciences Faculty of Biosciences P.O. Box 65, FI-00014 University of Helsinki, Finland
[email protected] Dr. NESHO CHIPEV, Associate Professor Central Laboratory of General Ecology Bulgarian Academy of Sciences 2, Gagarin Str., 1113 Sofia, Bulgaria
[email protected]
First published 2004 ISBN 954-642-224-X
© PENSOFT Publishers All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the copyright owner. Pensoft Publishers, Acad. G. Bonchev Str., Bl.6, 1113 Sofia, Bulgaria Fax: +359-2-870-45-08, e-mail:
[email protected] www.pensoft.net
Ecology of the Sofia City Page v
Contents
TABLE OF CONTENTS
ENVIRONMENT 3
The Sofia (Bulgaria) GLOBENET Sites: Description and Spatial Variation of the Landscape Mosaic Lyubomir Penev, Ivailo L. Stoyanov, Ivailo Dedov, Dimitar Dimitrov, Maria Grozeva, Vania Doichinova
25
Climate and Human Comfort of Sofia V. Andreev, Ch. Branzov, E. Koleva, A. Tsenkova, J. Ivantsheva, P. Videnov
55
Urban Climate and Air Quality – a Review Vassil M. Andreev
83
Air Quality of Sofia City – Meteorological Aspects V. Andreev, D. Syrakov, A. Tsenkova, J. Ivantsheva, P. Videnov
129
Water Quality of the Sofia Region Konstantin Tzankov, Plamen Ninov, Stefka Blaskova, Julia Kirova, Vessela Rajnova
165
Urban Planning and Ecological Problems of the Green Systems of Large Cities in Bulgaria: Sofia as a Case Study Atanas Kovatchev
PLANTS 173
Bryophytes in the city of Sofia Anna Ganeva
177
Mapping Air Quality Using a Lichen Diversity Index in Northeastern Sofia Dobri Ivanov & Andrey Kovatchev
185
The Vascular Flora of the City of Sofia Dimitar S. Dimitrov
209
Soil Moisture Regime in Forest Plantations in the Borisova Gradina Park and in the North Park in Sofia Petko Petkov & Georgi Antonov
213
Dendrological Structure of the Green Areas in Sofia Alexander Delkov & Raya Gateva
229
Environmental Problems for Trees in the Green System of Sofia Raya Gateva
249
Medicinal Plant Species and Edible Macromycetes in the City of Sofia – Dangerous for Man? Lilyana Yurukova
v
Ecology of the Sofia City Page vi
vi
L. Penev, J. Niemelä, D. J. Kotze & N. Chipev
257
Pathological and Entomological Problems of Trees in the Sofia City Green System Bojan Rossnev, Plamen Mirchev, Peter Petkov
269
Lead in the Atmosphere-soil-plant System in the Forest Park of Boris’ Garden – Sofia Svetla Bratanova-Doncheva
ANIMALS 281
Distribution Patterns of Nematode Communities in an Urban Forest in Sofia, Bulgaria Alexander Mladenov, Stela Lazarova, Vlada Peneva
299
Myriapoda (Chilopoda, Diplopoda) in Urban Environments in the City of Sofia Pavel Stoev
307
Spatial Variation in Terrestrial Gastropod Communities (Gastropoda, Pulmonata) along Urban-Rural Gradients in Sofia City, Bulgaria Ivailo Dedov & Lyubomir Penev
319
The Harvestmen Fauna (Opiliones, Arachnida) of the City of Sofia (Bulgaria) and its Adjacent Regions Plamen G. Mitov & Ivailo L. Stoyanov
355
Spiders from the Sofia Region. A Faunistic and Zoogeographical Analysis Andrej Antov, Stoyan Lazarov, Christo Deltshev and Gergin Blagoev
365
A Checklist of the Aphids (Homoptera, Aphidodea) from the City of Sofia Elena Tasheva-Terzieva
371
Spatial Structure of Carabid Beetle Assemblages along an Urban-Rural Gradient Ivailo L. Stoyanov & Lyubomir Penev
401
The Ground-Beetle (Coleoptera: Carabidae) Fauna of Sofia, Bulgaria: a Checklist Ivailo L. Stoyanov
417
A Faunistic Investigation of the Ants (Hymenoptera, Formicidae) in the City of Sofia Albena Lapeva-Gjonova & Maria Atanasova
423
Compound Nests and Mixed Colonies of Ant Species (Hymenoptera, Formicidae) in Sofia, Bulgaria Vera Antonova
429
Scuttle Flies (Diptera: Phoridae) from Urban and Suburban Areas in the Sofia Plain Mario S. Langourov
437
The Diurnal and Nocturnal Birds of Prey (Falconiformes and Strigiformes) Nesting in the Territory of Sofia Georgi P. Stoyanov, Dinyo T. Kyutchukov, Dobromir I. Domuschiev
451
Magpie (Pica pica) Population Density and Habitat Distribution in the City of Sofia Nevena T. Kamburova
Ecology of the Sofia City Page vii
Contents
vii
INTRODUCTION It is expected that in this year (2005), about half of the human population will be urban dwellers, and the proportion is increasing. Thus, urban areas – including green areas – are becoming increasingly important for humankind. In order to manage urban areas for the benefit of people and biodiversity more information is needed about ecological patterns and processes in cities. The present volume is a result of many years of investigations of the environment and biota of the city of Sofia, carried out by researchers from various academic institutions in Bulgaria. The book is intended to fulfill an obvious gap in the knowledge of the biota of Sofia. In spite of the various research projects carried out during the last decades, there is no volume published so far aiming to summarize the accumulated data. Such a gap is especially apparent when comparing Sofia to several other capitals or big European cities, such as Leipzig, Warsaw, Berlin, Prague, Rome and others. The idea of the book came about from the participants of the Bulgarian team of the GLOBENET project. The GLOBENET project was proposed by Jari Niemelä during a workshop in 1998 in Helsinki. GLOBENET aims at studying landscape changes in urban environment with both standardized sampling and data methods in various cities across the world using ground beetles (Coleoptera, Carabidae) as a model group. The rationale of GLOBENET is based on the “urban-to-rural gradient” concept proposed by Klausnitzer in 1982-83 and consequently by McDonnell and collaborators during the 1990’s. The concept assumes the existence, in virtually all big cities of the world, of a gradient, starting from heavily urbanized habitats within the city through suburban areas around the city, to rural habitats (natural or seminatural) in its vicinities. Shortly after the meeting in Helsinki, three international teams started sampling by standardized methods in Helsinki, Edmonton and Sofia. Furthermore, similar short-term projects have been carried out in Hiroshima (Japan), Debrecen (Hungary) and Hamburg (Germany). The habitats studied were forests – city parks, semi-natural forests in suburban areas and natural forests in rural areas. In each site, ground beetles were sampled using a standard number of pitfall traps. The samplings in Sofia resulted in a rich collection of other invertebrate groups. Therefore, the Bulgarian GLOBENET team extended the scope of research by inviting specialists in several groups of soil invertebrates, such as molluscs, millipedes and centipedes, spiders, ants and others to collaborate in the research. The aim was to study how the different organisms and their assemblages react to urbanization. During the course of research, we realized that a useful interpretation of the results is hardly possible without using data from other research projects in urban ecology. Therefore we invited several Bulgarian scientists from different academic institutions – zoologists, botanists, ecologists, climatologists, forest researchers and environmentalists – to summarize their knowledge on the environment and biota of the city of Sofia. The outcome of this multi-institutional cooperation is summarized in the present volume. The institutions involved are: Central Laboratory of General Ecology of the Bulgarian Academy of Sciences (BAS), Institute of Botany (BAS), Institute of Zoology (BAS), Forest Research
Ecology of the Sofia City Page viii
viii
L. Penev, J. Niemelä, D. J. Kotze & N. Chipev
Institute (BAS), Institute of Hydrology and Meteorology (BAS), Sofia University St Kliment Ohridski and the University of Forestry, Sofia. The book consists of three parts – Environment, Plants and Animals. The papers in the first part aim to describe the environment of Sofia in order to provide a basis for present-day and future research on the biota. A research team from the Institute of Hydrology and Meteorology of the Bulgarian Academy of Sciences provided an extensive review on the climate, air and water quality of Sofia. An interesting view on the impact of climate on human comfort is discussed in the paper by Andreev et al. The Green System of Sofia, its history, development and especially the problems generated by conflicts between land owners and municipality during the period of transition to market economy are briefly presented in the paper by Atanas Kovachev. The Sofia GLOBENET Sites are described in a separate paper by Penev et al. in order to provide the environmental background for several papers in this volume dealing with variation in species composition and assemblages of soil invertebrates along urban-to-rural GLOBENET gradients. The botanical part consists of 9 papers, ranging from inventories (bryophytes, vascular plants) to studies on the soil moisture regime and lead transformation in city parks. Of special interest is the map of air quality produced by using the Lichen Diversity Index discussed in the paper of Ivanov & Kovatchev. The potential dangers to man of using medicinal plants and mushrooms collected within the city are studied in a paper by Yurukova. The tree vegetation of Sofia is of special interest to several academic institutions and three papers are devoted to its dendrological structure (Delkov & Gateva), impact of pollution on trees (Gateva) and pathological and entomological issues of the tree vegetation in urban environment (Rosnev et al.). The third part of the book deals with several animal groups. Eight papers summarize data sampled in the GLOBENET sites – nematodes (Mladenov et al.), myriapods (Stoev), terrestrial gastropods (Dedov & Penev), opilionids (Mitev & Stoyanov), spiders (Antov et al.), carabid beetles (Stoyanov & Penev; Stoyanov) and ants (Lapeva-Gjonova & Atanasova). Besides, two more papers present inventories of the urban fauna of aphids (Tasheva) and scuttle-flies (Langurov). Compound nests and mixed colonies of ants in city parks are studied by the youngest member of the GLOBENET team – Vera Antonova. Finally, birds as an important element of the biodiversity of the city are discussed in the papers of Stoyanov et al. and Kamburova. This volume aims to summarize present-day knowledge on the environment, ecology and threats to life in the city of Sofia, offered by various academic teams. Therefore, one cannot expect that the present book will offer the final word on the biota and environment of Sofia. The Editors hope that the conclusions drawn in the different studies can be used in urban planning and decision-making processes. We also hope that this first volume will be a solid basis and stimulus for further research on the ecology of Sofia. The Editors
Ecology of the Sofia City CyanMagentaYellowBlack Page 1
The Sofia (Bulgaria) GLOBENET Sites: Description and Spatial Variation of the Landscape Mosaic
Environment
1
Ecology of the Sofia City CyanMagentaYellowBlack Page 2
Lyubomir Penev et al.
This page intentionally left blank
2
Ecology of the Sofia City CyanMagentaYellowBlack Page 3
The Sofia (Bulgaria) GLOBENET Sites: Description and Spatial Variation of the Landscape Mosaic
The Sofia (Bulgaria) GLOBENET Sites: Description and Spatial Variation of the Landscape Mosaic Lyubomir Penev 1, Ivailo L. Stoyanov 1, Ivailo Dedov 1, Dimitar Dimitrov 2, Maria Grozeva 3, Vania Doichinova 3 1
2
Central Laboratory of General Ecology, 2 Yurii Gagarin Str., 1113 Sofia National Museum of Natural History, 1 Tzar Osvoboditel Blvd., 1000 Sofia 3 Forest Research Institute, 132 St. Kliment Ohridski Blvd., 1756 Sofia
ABSTRACT The present study summarizes all the main physical and phytological features of the sampling sites, examined during the Sofia GLOBENET Project. Based on these data, an attempt for classifying the structure of the investigated landscape mosaic has been made. As a result, the presence of a strong urban-rural gradient, with no distinct intermediate “suburban” zone, was demonstrated. The gradient also showed a naturally hierarchic sub-structure. The prominent landscape descriptors that determine the described landscape mosaic patterns were identified – some of the most important include: origin of the forest, altitude, patch area, litter humidity, and built-up area. KEY WORDS Urban habitats, urban-rural gradient, GLOBNET, Sofia, Bulgaria.
INTRODUCTION The spatial variation of biotic communities in urban environments became recently a field of study that enjoy a prime interest of biologists, landscape ecologists, and specialists in urban planning. The increasing number of papers and monographs treating species’ behavior in, and biotic communities’ responses to urbanization processes, is evidence for the growing need to investigate how the specific conditions of the big cities affect the environment, and how this environment can be adapted in order to ensure both the protection of biodiversity and improving the living conditions for the people. For an extensive summary on the role of soil ecology (and particular soil zoology) in this process see Whiteley (1994). Urbanization seems to have quite similar characteristics in different parts of the world. Although this statement appears intuitive, some evidence exists that the landscape structure L. Penev, J. Niemelä, D. J. Kotze & N. Chipev (Eds.) 2004. Ecology of the City of Sofia. Species and Communities in an Urban Environment, pp. 3-24 © PENSOFT Publishers Sofia-Moscow
3
Ecology of the Sofia City CyanMagentaYellowBlack Page 4
Lyubomir Penev et al.
affected by urbanization retains its specificity at different places around the globe (Reynaud & Thioulouse, 2000), while globalization leads to a “uniformity” in the look of big cities around the world. At the same time, urbanization exists in virtually all floristic/faunistic regions of the Globe, affecting local flora and fauna of rather different features, such as species composition and diversity, age of association, adaptiveness and so on. Hence, the idea of investigating the impact of “a common factor onto different floras/faunas in different ecological/historical conditions” or “intercontinental community convergence” (Schoener, 1986) finds an appropriate application in times of global urbanization. One of the first steps in this direction was the establishment of the GLOBENET project, aimed at a survey of the common anthropogenic impacts on biodiversity on a global scale using carabid beetles, launched in 1997 by Jari Niemelä and co-workers at the first GLOBENET Workshop in Helsinki, Finland (Niemelä et al., 2000; Niemelä & Kotze, 2000). The basic concept lies in the presumption that monitoring landscape changes caused by urbanization in similar habitats (forests) over different geographical regions, using one and the same focal taxonomic groups (ground beetles – Carabidae, Coleoptera), and employing the same sampling methodology, would lead to useful conclusions about the reaction of biotic communities anthropogenic pressure (Niemelä & Kotze, 2000). The idea has quickly found supporters, and the work started the year after the GLOBENETWorkshop in three cities of Europe and North America – Helsinki (Finland), Sofia (Bulgaria), and Edmonton (Canada). The first results were reported at the Second GLOBENET Meeting in Barziya (Bulgaria) in May 1999. The increased interest in the project encouraged the Sofia team to extend the work towards incorporating studies on soil and floristic components of the GLOBENET sites, as well as inviting specialists in some important soil mesofauna groups, such as gastropods, spiders, myriapods, opiliones, ants, nematodes and others. The present paper aims at: 1) presenting a detailed physical and floristic description of the Sofia GLOBENET Sites, 2) describing the methods used in measuring these site characteristics, 3) analyzing their joint variation in order to describe, and quantify the hypothesized gradient, and 4) classifying the investigated landscape mosaic.
MATERIALS AND METHODS General information on the relief, climate, and soils in the investigated area were gathered from the literature (Vaptsarov et al., 1997; Bluskova et al., 1983; Ninov, 1997). The area, distance from the city center, and exposition of the sampling sites were directly estimated from maps of the city and its surroundings (Smilenova et al., 1996). The variable “built-up area”, designed to provide an alternative estimate of urbanization, was also estimated from the maps and is expressed as a percentage of the area covered with buildings within a 2 x 2 km square centered at the respective study plot. Altitude was measured with a portative altimeter. Additionally, data on the tilt, origin, age and height of the tree stands, were obtained from the database of the Forest Committee. The canopy cover was estimated visually. Further, the “number of trees” and “diameter of trees” variables were obtained by averaging the count of trees and the tree 4
Ecology of the Sofia City CyanMagentaYellowBlack Page 5
The Sofia (Bulgaria) GLOBENET Sites: Description and Spatial Variation of the Landscape Mosaic
diameter (measured at breast height) within five 10 x 10 m square samples at each sampling plot. Leaf-litter thickness is an average of 140 randomly placed measurements. The ordinal variables “grass cover”, “humidity of leaf-litter”, and “humidity of soil” were estimated (140 per site) subjectively on a scale from 0 (absent) to 5 (very strong). To estimate nonparametrically the “location” of each sampling site (based on the latter three variables) for the purposes of further analysis, we calculated the rank-sums of each variable and subsequently ranked these to obtain the values shown in Table 3 and entered into the multivariate analyses. The anthropogenic soils were classified after Gencheva (1995), while for the non-antropogenic soils we used the classification of FAO-UNESCO (1988). The chemical and mechanical properties of soil at the sampling sites were obtained as follows: pH was potentiometrically estimated using a “Practronic MV98”; organic carbon and humus were assessed with the Turin method (e. g. Arinushkina, 1970); nitrogen contents of the soil was estimated as Total Kjeldahl Nitrogen (TKN) (Isaac & Johnson, 1976); and the mechanical composition of the soil was determined following the method of Kachinsky (1958). In the numerical analyses the data presented in Tables 1-4 (except for the variable “ground” which had some categories of insufficient frequency) were entered into a Principal Component Analysis after centring and normalizing. For identifying of naturally hierarchical clusters of rows (i. e. sampling sites) in the analyzed matrix, a repeated “K-means” clustering procedure (based on the algorithm of Hartigan & Wong (1979)) with K varying from 2 to 5 clusters was employed on the first two principal components. All the computations were performed in the R language environment (Ihaka & Gentlemen, 1996).
RESULTS One of the main criteria used to select the sampling sites was the prevalence of more or less the same type of forest – in our case broad-leaved forests dominated by oak (Quercus L.). One Table 1. Physical characteristics of the Sofia GLOBENET sites. For abbreviation description and disposition of sites see Fig. 1. Code Area [ha] UI UII UIII UIV SI SII SIV RI RII RIII RIV
1.9 1.2 3.7 1.2 1.9 8.0 13.7 18.0 15.2 7.0 15.0
Exposition
Altitude [m]
E-SE W-WS N-NE S-SE N N N, NW-W NE NE N W
550 525 575 570 650 900 950 850 920 800 1000
Slope [degree] Distance [km] 4 4 4 4 17 14 17 24 14 7 18
3.3 3.8 5.6 3.7 13.2 12.0 11.3 14.5 13.5 29.0 10.2
Built area [%] 71 64 35 56 13 6 27 0.2 0 0.3 0.3
5
Ecology of the Sofia City CyanMagentaYellowBlack Page 6
Lyubomir Penev et al.
Fig. 1. The Sofia GLOBENET sites: U1 – Knyaz Borissovata Gradina; U2 – Zapaden Park; U3 – Severen Park; U4 – Loven Park; S1 – SE of German Village; S2 – locality “Vilite” (Lyulin Mt.); S4 – SE of Vladaya Village (Vitosha Mt.); R1 – S of Germanski Monastery (Lozenska Mt.); R2 – N of locality “Manastirski Polyani” (Lyulin Mt.); R3 – S of Drenovo Village; R4 – locality “Tikhiya Kat” (Vitosha Mt.)
reason for the choice of similar habitats along the urban-rural gradient was to minimize the additional variation of different habitat types on the soil mesofauna. Another reason that contributed to the elaboration of the mentioned sampling strategy, was the fact that this habitat type have been the most significantly transformed during (at least) the last 300 years, being naturally dominant in the investigated region for centuries (see Stoyanov, 1937). Eleven sites, situated along the urban-to-rural gradient (having its urban end in the city parks of Sofia, and spreading radially through the suburban zone towards rural areas – respectively in southeastern (1), southwestern (2), northern (3), and southern (4), directions) were selected (Fig. 1). Based upon our convention about the sampling design, and limited by the availability of appropriate forest habitats, each gradient was preliminary stratified into three zones, reflecting the varying degree of urbanization, namely “urban” (U), “suburban” (SU), and “rural” (R). As already mentioned, due to the complete absence of appropriate forest habitats northwards of the city of Sofia, the northern gradient lacks the intermediate (SU) zone (Fig. 1).
6
Ecology of the Sofia City CyanMagentaYellowBlack Page 7
The Sofia (Bulgaria) GLOBENET Sites: Description and Spatial Variation of the Landscape Mosaic
Physical description of the sites Location and relief The investigated area is situated in West Bulgaria and includes the Sofia Kettle and foothills of the surrounding mountains – Stara Planina, Vitosha, Sredna Gora, and Lyulin, up to 1000 m a. s. l. The association of the mountains is completely tectonic. Geomorphologically all sites belong to the South Bulgarian province of the Trans-Balkan kettle subregion (Vapzarov et al., 1997). Climate According to climatic regioning, the investigated region pertains to the temperate-continental climatic region of Bulgaria. This climatic region is characterized by warm summers, cold winters, the annual air temperatures have a significant amplitude, spring-summer maximum and winter minimum of rainfalls, and a relatively steady snow cover with altitude-related duration (Velev, 1997). Generally, the climate of the city of Sofia is comparatively milder than the surrounding regions and is characterized by higher mean temperatures of the coldest month and a cooler summer (Bluskova et al., 1983). Soils The leached chromic livisols, present in sites Knyaz Borisova Gradina (U1) and Zapaden Park (U2), are characterized by a rich humus horizon with 5-6 % humus (Tab. 2). Their structure is grainy, turning downwards into a prismatic one, with heavier mechanical composition (31-54% clay content), hence higher hygroscopic moisture. pH varies between 4.6 and 6.5. The color of the soils varies from leached black to dark brown with low differentiation along the soil profile. The water tightness is low and the water-air regime is worsened. In comparison to the above described city parks, the soils in Loven Park (U4) are more acid, with lighter mechanical composition in the upper layer and can be determined as transitional form to the maroon chromic livisols. The soils in Severen Park (U3) possess typical anthropogenic features. Their chemical Table 2. Soil characteristics of the Sofia GLOBENET sites. Code
Ground
pH
UI UII UIII UIV SI SII SIV RI RII RIII RIV
Chromic livisols Chromic livisols Urbogenic anthroposoils Maroon chromic livisols Leached maroon forest soils Leached maroon forest soils Leached maroon forest soils Leached maroon forest soils Maroon soils Leached maroon forest soils Leached maroon forest soils
6.50 5.85 7.05 5.30 6.65 4.65 5.50 6.60 4.90 5.75 5.10
Humus [%] Mechanical composition 5.83 5.06 10.67 6.38 7.73 3.67 9.61 6.33 4.97 7.95 8.16
Medium Heavy Heavy Medium Light Light Medium Light Light Heavy Medium
N [%] 0.18 0.16 0.21 0.14 0.27 0.15 0.26 0.19 0.16 0.21 0.27
7
Ecology of the Sofia City CyanMagentaYellowBlack Page 8
Lyubomir Penev et al.
reaction is neutral to low-alkaline (pH 6.9-7.05), the humus content is high and weakly differentiated along the profile. The mechanical composition is heavy, and according to Gencheva (1995), these soils can be determined as terri-cumulic, urbogenic anthroposols. The foothills of the surrounding mountains are covered by leached maroon forest soils. The sites at Lozenska Mt. (German Village, S1) and Lyulin Mt. (Vilite, S2) are characterized by a less strong profile (70-80 cm), heavier mechanical composition with stronger differentiation along the profile. The chemical reaction of the soil solution is acid to low-acid (pH approximately between 4 and 6), the humus content is low. The soils in the region of Vladya Village (S4) are deeper (over 90 cm), in comparison to the previously mentioned sites, with a lighter mechanical composition, a more acid reaction (pH 4–5), and rather diverse with respect to the humus content (0.2 to 9 %). With an increase in altitude up to 700–800 m a. s. l., thinner maroon soils become prevailing at sites “R3” and “R4”. The site near the German Monastery (R1) differs by the low alkaline reaction (pH 6–8) of its soils, due to the carbonate basic rocks in that area. In the negative forms of the relief, as it is the case at site “R2” (situated on former terraces), a kind of meadow maroon soils are present. They are characterized by a strong (100 cm) and rather differentiated soil profile, with low-acid reaction (pH about 6.5). Besides, these soils are characterized by a low humus content (about 3 %) and are rather influenced by the higher level of underground water. Biological description of the sites Plant communities According to the floristic regioning of Bulgaria (Jordanov, 1966), the study sites are situated in four different floristic regions: 1) Sofia Region – all urban sites (U1-4); 2) Western Sredna Gora Region – “S1” and “R1”; 3) Vitosha Region – “S2”, “R2”, “S4”, “R4”; 4) Western Stara Table 3. General features of the plant communities of the Sofia GLOBENET sites. Code Origin
UI UII UIII UIV SI SII SIV RI RII RIII RIV
8
Age Canopy Height Number Diameter Thickness Grass- Humidity Humidity [years] [%] [m] of trees of trees of leaf- cover of leaf- of the [100m-2] [cm] litter [cm] litter soil
Artificial Artificial Artificial Artificial Artificial Secondary Secondary Secondary Secondary Secondary Secondary
90 50 60 90 40 60 50 20 50 50 70
85 90 70 90 80 85 85 60 85 80 80
22 15 30 26 18 18 15 20 17 17 25
15,4 16,6 8,4 17,6 16,4 22,4 11,2 14,6 25,4 16,2 10
18.4 20.0 23.3 19.4 16.0 14.1 18.9 16.5 11.5 17.0 20.6
2,2 3,7 2,8 3,8 4,0 2,8 3,3 3,1 3,0 3,3 2,9
9 1 4 3 2 10 8 5 7 6 11
9 6 10 11 7 2 5 1 3 4 8
3 10 11 9 6 5 7 1 2 4 8
Ecology of the Sofia City CyanMagentaYellowBlack Page 9
The Sofia (Bulgaria) GLOBENET Sites: Description and Spatial Variation of the Landscape Mosaic
Planina Region – “R2”. This reflects local specificity of the floristic composition of the sites (Appendix I) caused by both local physical peculiarities of the site and its geographical position. All sampling sites belong to the zone of hornbeam-oak woods ranging, in Bulgaria, from 600-700 to 1000-1200 m a. s. l. (Bondev, 1991). Sites “S2”, “R2”, “S4” and “R4” approach the beech-woods belt. The main criterion for selecting the GLOBENET sampling sites was the presence of broadleaved forests, dominated by oak-tree species. In the city parks, the introduced mesophilous oak species Quercus rubra prevails. In the natural and semi-natural habitats the commonest oak species are Q. cerris (S4, R4), Q. pedunculiflora (U1) and Q. polycarpa (S2, R2) (Appendix I). The herb layer is dominated by Dactylis glomerata and Poa nemoralis, which seem to be the most common herb species and are recorded from all sites. Some species are present only in urban environment (Bromus inermis and Elymus hispidus), while other are found only in suburban and rural habitats (all Carex spp. and Danthonia alpina). Some species occur in most of the suburban and rural habitats but are absent in the urban environment (Festuca heterophylla at “S4”, “R1”, “R2”, “R4”, and Melica uniflora at “S2”, “S4”, “R1”, “R2”). There are also obvious geographical peculiarities in the distribution of certain species, i. e. such characteristic for the Vitosha Floristic Region only (S4, R4 – Agrostis capillaris, Arrenatherum elatius, Briza media, Bromus ramosus, Calamagrostis arundinacea, Elymus caninus, and Luzula luzuloides). The floristic specificity of the rural habitats of the Western Stara Planina Region are marked by the presence of Festuca valesiaca and Luzula campestris, the Western Sredna Gora Region – by Luzula forsteri and Poa pratensis. The sampling plot “U1” is a mixed deciduous-coniferous forest, dominated by Quercus pedunculiflora (the plant formation is Querceta pedunculiflorae). The forest-habitat “U2” is dominated by Quercus rubra (the plant formation is Querceta rubrae). The forest at sampling-site “U3” may be subdivided into two homogeneous sub-sites: the first one is a deciduous forest dominated by Quercus rubra (formation Querceta rubrae), the second one is a mixed deciduous-coniferous forest, dominated by Populus nigra (formation Populeta nigrae). At “U4” the mixed deciduous forest is dominated by Quercus rubra (formation Querceta rubrae). Sampling site “S1” is a plantation almost entirely consisting of Quercus rubra (formation Querceta rubrae). The next suburban sampling plot (S2) is a mixed deciduous forest dominated by Quercus polycarpa, the other common tree species is Carpinus betula (formation Querceta polycarpae). The dominating tree at study site “S4” is Quercus cerris (Querceta cerris). The rural sampling site “R1” is a mixed deciduous forest with Quercus sessiliflora and Carpinus betulus (formation Querceta sessiliflorae). Site “R2” is a mixed deciduous forest dominated by Carpinus betulus and Quercus polycarpus (formation Querceta ploycarpae). The highly isolated forest, where sampling site “R3” is located, is dominanted by Quercus cerris. Finally, site “R4” is a Quercus cerris forest (Querceta cerris). For a full plant species composition of the sampling sites, refer to Appendix I.
DISCUSSION Looking at the overall data structure that emerged from the multivariate analyses of our datasets (Figs. 2-10), the main landscape gradient is clearly visible in most of the cases. When analyzing the plant cover of the sampled landscape, the clearest support of our initial working 9
Ecology of the Sofia City CyanMagentaYellowBlack Page 10
Lyubomir Penev et al. d=2
U
R
S
Fig. 2. PCA ordination of the Sofia GLOBENET sites (grouped by zone), based on species composition of grasses. d=5
R
US
Fig. 3. PCA ordination of the Sofia GLOBENET sites (grouped by zone), based on species composition of herbs.
10
Ecology of the Sofia City CyanMagentaYellowBlack Page 11
The Sofia (Bulgaria) GLOBENET Sites: Description and Spatial Variation of the Landscape Mosaic d=2
U
R S
Fig. 4. PCA ordination of the Sofia GLOBENET sites (grouped by zone), based on species composition of shrubs. d=2
R
S
U
Fig. 5. PCA ordination of the Sofia GLOBENET sites (grouped by zone), based on species composition of trees.
11
Ecology of the Sofia City CyanMagentaYellowBlack Page 12
Lyubomir Penev et al.
hypothesis (i. e. the urban-to-rural gradient) is the tree stratum of the vegetation communities (Fig. 5). As plant size decrease (respectively its edification role), through the series trees-shrubsherbs-grasses (Figs. 5, 4, 3, 2) the presumed urban-to-rural-gradient becomes less prominent. When including all the vegetation layers in the analysis (Fig. 6), the urban zone becomes most distinct and clearly separates from the non-urban zones. Here we are speaking of “non-urban” zone mainly to translate our findings to the framework of the implemented urban-suburbanrural sampling design, and to refer to the strongly overlapping “rural” and “suburban” sites. When including all the measured environmental variables (with the vegetation data), the picture of the urban-to-rural gradient becomes even clearer (Fig. 7). The urban zone is again clearly different form the other two, which in turn are more or less intergrading. When considering the relationships between the measured environmental properties of the sampling plots, a not so complex structure, although of strongly intercorrelated variables, in the investigated landscape becomes evident (Fig. 8). The principal component analysis reveals the existence of one prominent factor summarizing about 41% of the total variation within the landscape. The environmental variables that are most strongly correlated with this first principal component of the data are: forest origin, altitude, and area (positively correlated), and litter humidity, builtup area, tree diameter (negatively correlated). The second principal component summarizes about half the variation of the data, and is most strongly correlated with tree number/100 m, canopy, and vegetation PCA axis 1 (positively), and with humus contents, nitrogen contents, and tree diameter (negatively). Finally, our attempt to classify the observed structures revealed a strong and (most important) hierarchical organization (Fig. 10) of the sampled landscape
d=5
R
S
U
Fig. 6. PCA ordination of the Sofia GLOBENET sites (grouped by zone), based on all plant species.
12
Ecology of the Sofia City CyanMagentaYellowBlack Page 13
The Sofia (Bulgaria) GLOBENET Sites: Description and Spatial Variation of the Landscape Mosaic d=2
U S R
Fig. 7. PCA ordination of the Sofia GLOBENET sites (grouped by zone), based on environmental variables. d = 0.5 tree.n
canopy veg.pca.1
build tree.age
litter.th veg.pca.2
dist gras s .cover
ter.hum m ech.com p s oil.hum
fores t.o alt s lope area
pH tree.height
trees .diam
N hum us
Fig. 8. PCA ordination of the environmental variables measured at the Sofia GLOBENET sites (F1=40.5%, F2=19.6%).
13
Ecology of the Sofia City CyanMagentaYellowBlack Page 14
Lyubomir Penev et al. d=2
S2 SII UIV 4 U
R2 II
U II2
U U 1I
SI1 S
RIII 3 R
R R 1I S4 SIV
UIII 3 U RIV 4 R
Fig. 9. PCA ordination of the Sofia GLOBENET sampling sites, based on environmental data (the factorial plane summarizes 60.1% of the total dataset variability). d=2
d=2
1
1
2 2 3
K=2 clusters
K=3 clusters d=2
d=2
2
3
3
2
4 4 1 1
K=4 clusters
5
K=5 clusters
Fig. 10. Sequential “K-means” (starting from K=2 to K=5) clustering of the PCA ordination of the Sofia GLOBENET sites, based on environmental data.
14
Ecology of the Sofia City CyanMagentaYellowBlack Page 15
The Sofia (Bulgaria) GLOBENET Sites: Description and Spatial Variation of the Landscape Mosaic
structures (Fig. 9). The sequential “K-means” partitioning procedure yielded homogeneous clusters with virtually no interchangeable cluster-members. For example at the 2-cluster level of the procedure, the urban zone clearly opposites the non-urban one. At the subsequent 3cluster level, the urban zone remains compact, while the non-urban splits in two (mainly due to the isolated position of sites at Lyulin Mt. – S2 and R2). The urban cluster parted when we have repeated the clustering with K=4. The most peculiar site within this group appears to be Severen Park (U3). The remaining partitioning subdivided the greater non-urban cluster, hinting that at least partially the site-group containing “S1”, “S4” and “R4” may be considered a transitional zone in the sense of our working hypothesis (Fig. 10).
CONCLUSIONS Based on the above considerations, the following conclusions may be drawn: 1) A clear urban-to-rural gradient was demonstrated. 2) Our working hypothesis about the presence of a sharply delimited and self-contained “suburban” zone was not supported by the analyses. 3) The classification of the observed gradient structure showed the existence of a natural and hierarchic sub-structure. 4) We revealed the prominent environmental factors determining the investigated landscape structure. These are: origin of forest, altitude, patch area, litter humidity, built-up area, tree diameter, tree number, canopy, general vegetation variability (as expressed by the first PCA axis of a separate analysis), humus and nitrogen content of the soil, tree diameter.
ACKNOWLEDGEMENTS This work was partly supported by the National Scientific Fund (grant B-808/1998). We thank Dr. D. J. Kotze for final editing the English text.
REFERENCES Arinushkina E.V. 1970. Manual for Chemical Analysis of Soils. 2nd Ed., State University, Moscow, 487 pp. [In Russian] Bluskova, D., Zlatkova, L., Lingova, S., Modeva, Z., Subev, L. & Teneva, M. 1983. Climate and Microclimate of Sofia. Sofia, Publishing House of the Bulgarian Academy of Sciences, 153 pp. (in Bulgarian). Bondev, I. 1991. Vegetation of Bulgaria. Sofia, “St. Kliment Ohridsky” Publishing House. 183 pp. FAO-UNESCO 1988. Soil Map of the World. Revised Legend, Rome. Gencheva, S. 1995. Classification and peculiarities of the anthropogenic soils. PhD Thesis, Sofia, University of Forestry, 48 pp. (in Bulgarian). Hartigan, J.A. & Wong, M.A. 1979. A K-means clustering algorithm. Applied Statistics 28: 100–108. Ihaka, R. & Gentleman, R. 1996. R: A Language for Data Analysis and Graphics. Journal of Computational and Graphical Statistics 5(3): 299-314.
15
Ecology of the Sofia City CyanMagentaYellowBlack Page 16
Lyubomir Penev et al. Isaac, R. A. & Johnson, W. C. 1976. Determination of total nitrogen in plant tissue, using a block digestor. J. Assoc. Off. Anal. Chem. 59: 98–100. Jordanov, D. 1966. Flora Bulgarica, vol. III. Sofia, Publishing House of the Bulgarian Academy of Sciences, 635 pp. (in Bulgarian). Kachinsky, N.A. 1958. Mechanical and micro aggregate composition of soils and methods for their investigation. MGUPress, Moscow, 192 pp. (in Russian). Niemelä, J., Kotze, J., Ashworth, A., Brandmayr, P., Desender, K., New, T., Penev, L., Samways, M., & Spence, J. 2000. The search for conmmon anthropogenic impacts on biodiversity: a global network. Journal of Insect Conservation 4: 3-9. Niemelä, J. & Kotze, J. 2000. GLOBENET: the search for common anthropogenic impacts on biodiversity. In: Natural History and Applied Ecology of Carabid Beetles, Proceedings of the IX European Carabidologists Meeting, Cosenza, Italy, 26-31 July, 1998. (Brandmayr, P., Lövei, G., Brandmayr, T. Z., Casale, A., Vigna Taglianti, A., Eds.), pp. 187-196. Sofia-Moscow, Pensoft Publishers. Ninov, N. 1997. Soil-geographic regioning. – In: Geography of Bulgaria. (Yordanova, M. & Donchev, D., Eds.), pp. 251-257. Sofia, “Prof. Marin Drinov” Academic Publishing House. (in Bulgarian). Reynaud, P.A. & Thioulouse, J. 2000. Identification of birds as biological markers along a neotropical urbanrural gradient (Cayenne, French Guiana), using co-inertia analysis. Journal of Environmental Management 59: 121-141. Schoener, T. W. 1986. Mechanistic approaches to community ecology: a new reductionism? Am. Zool. 26: 81-106. Smilenova, L., Bozhilova, M., Starikova, L. (Eds.) 1996. Sofia/Vitosha maps (1:20000/1:50000). Sofia, Kartografia, 47 pp. (in Bulgarian). Stoyanov, N. 1937. The vegetational relationships in the Sofia kettle. Agronomo-lesowudski Godishnik 1: 1-40. (in Bulgarian). Vaptzarov, I., Alexiev, G. Vlaskov, V. 1997. Geomorphologic regioning, In: Geography of Bulgaria. (Yordanova, M. & Donchev, D., Eds.), pp. 103-105, Sofia. “Prof. Marin Drinov” Academic Publishing House. (in Bulgarian). Velev, S. 1997. Climatic regioning, In: Geography of Bulgaria. (Yordanova, M. & Donchev, D., Eds.), pp. 127130. Sofia, “Prof. Marin Drinov” Academic Publishing House. (in Bulgarian). Whiteley, D. 1994. The state of knowledge of the invetrebrates in urban areas in Britain with examples taken from the city of Sheffield. Memorabilia Zoologica 49: 207-220.
16
Ecology of the Sofia City CyanMagentaYellowBlack Page 17
The Sofia (Bulgaria) GLOBENET Sites: Description and Spatial Variation of the Landscape Mosaic Appendix I. Species composition of the plant communities of the Sofia GLOBENET sites.
Trees Acer campestre L. Acer hyrcanum Fisch. et C.A. Mey. Acer platanoides L. Acer pseudoplatanus L. Acer tataricum L. Betula pendula Roth Carpinus betulus L. Cornus mas L. Cornus sanguinea L. Coryllus avellana L. Fagus sylvatica L. Fraxinus excelsior L. Fraxinus ornus L. Juglans regia L. Larix decidua Mill. Quercus cerris L. Quercus pedunculiflora C. Koch Quercus polycarpa Schur Quercus rubra L. Quercus sessiliflora Salisb. Pinus nigra Arn. Pinus strobus L. Populus nigra L. Populus tremula L. Prunus avium L. Prunus cerasifera Ehrh. Pseudotsuga glauca Mayr Pyrus pyraster Burgsd. Robinia pseudoacacia L. Sambucus nigra L. Sophora japonica L. Sorbus aucuparia L. Sorbus torminalis (L.) Tilia cordata Mill. Tilia tomentosa Moench Ulmus minor Mill. Shrubs Chamaecytisus hirsutus (L.) Clematis vitalba L. Crataegus monogyna Jacop. Euonymus europaeus L. Euonymus verrucosus Scop
U1
U2
+
+
U3
U4
S1
S2
S4
+
+
+
R1
R2
R3
+
R4 +
+ +
+ + +
+ + +
+
+
+
+ +
+ +
+
+
+ +
+
+ +
+
+
+
+
+ +
+ +
+
+ + +
+
+
+
+ +
+
+
+
+
+
+
+ + + +
+
+ + +
+
+
+
+ +
+
+ +
+
+ +
+ + +
+
+ +
+ +
+ + +
+
+ +
+
+
+ +
+
+
+
+
+
+
+
+ +
+
+
+
+ +
+ +
+
+
+ +
+ +
+ +
+
+ +
+
+ + +
+
+
+ +
+
+ +
+
17
Ecology of the Sofia City CyanMagentaYellowBlack Page 18
Lyubomir Penev et al. Appendix I. Continued. U1 Hedera helix L. + Juniperus communis L. Ligustrum vulgare L. + Lonicera tatarica L. Mahonia aquiflora (Pursh.) Philadelphus coronarius L. Prunus spinosa L. Rosa canina L. + Rosa dumalis Bechst. Rosa jundzillii Besser Rubus caesius L. Rubus idaeus L. Rubus macrophyllus Weihe et Nees Symphoricarpos racemosus Mich. Viburnum lantana Hemsl. Viburnum opulus L. Grasses Agrostia capillaris L. Arrenatherum elatius (L.) Artemisia vulgaris L. Atriplex patula L. Brachypodium sylvaticum (Huds.) Briza media L. Bromus inermis Leyss. Bromus ramosus Huds. Calamagrostis arundinacea (L.) Carex cuprina Carex divisa Huds. Carex silvatica L. Dactylis glomerata L. Danthonia alpina Vest. Elymus caninus (L.) Elymus hispidus (Opiz.) Festuca heterophylla Lam. Festuca valesiaca Schleich. Galium odoratum (L.) Lapsana communis L. Luzula campestris (L.) Luzula forsteri (Sm.) Luzula luzuloides (Lam.) Melica uniflora Retz. Poa nemoralis L. Poa pratensis L.
18
U2
U3
+ +
+ +
U4
S1
+
+
+
+
S2
S4
R1
+ +
R2
R3
R4
+
+
+
+
+ + + +
+ +
+ + +
+ +
+
+ + +
+ + +
+
+ +
+ + + +
+ + +
+
+ +
+
+ +
+
+
+
+
+
+
+ + + +
+
+
+
+ +
+ +
+
+
+ +
+ + + +
+
+
+ +
+ + +
+ + +
+ +
+
Ecology of the Sofia City CyanMagentaYellowBlack Page 19
The Sofia (Bulgaria) GLOBENET Sites: Description and Spatial Variation of the Landscape Mosaic Appendix I. Continued. U1 Herbs Achillea millefolium L. Agrimonia eupatoria L. Ajuga genevensis L. Alliaria petiolata (Bieb.) Anthriscus cerefolium (L.) Arabis recta Vill. Arabis turrita L. Arctium lappa L. Aremonia agrimonoides (L.) Arum maculatum L. Ballota nigra L. Betonica officinalis L. Buglossoides purpurocoeruleus (L.) Campanula bononiensis L. Campanula persicifolia L. Cardamine bulbifera (L.) Centaurea jacea L. Centaurea phrygia L. Centaurea rutifolia S.S. Cerastium caespitosum (L.) Cichorium intybus L. Clinopodium vulgare L. Convallaria majalis L. Corydalis slivenensis Vel. Cruciata glabra (L.) Cruciata laevipes (L.) Dianthus armeria L. Digitalis viridiflora Lindl. Echinops sphaerocephalus L. Epilobium hirsutum L. Epipactis helleborine (L.) Erithronium dens canis L. Euphorbia amygdaloides L. Euphorbia cyparissias L. Filipendula vulgaris Moech. Fragaria vesca L. Galim schultesii Vest. Geranium robertianum L. Geranium sanguineum L. Geum urbanum L. Glechoma hederacea L. Helleborus odorus Waldst et Kit.
U2
U3
U4
S1
S2
S4
R1
R2
R3
R4
+ + +
+ +
+ +
+ +
+
+
+
+ +
+ + +
+ +
+ +
+ + + +
+ + + + +
+ +
+ +
+ +
+ +
+
+
+
+
+ + + + +
+ + + +
+
+ +
+
+
+
+
+ + +
+
+
+ + +
+ +
+
+
+
+
+ + + + + + + + +
19
Ecology of the Sofia City CyanMagentaYellowBlack Page 20
Lyubomir Penev et al. Appendix I. Continued. U1 Hieracium laevigatum Willd. Hieracium umbellatum L. Hypericum perforatum L. + Knautia drymeja Henff. Lamium purpureum L. Lapsana communis L. Lathyrus laxiflorus (Desf.) Lathyrus niger L. Lathyrus vernus L. Lentodon asper (W.K.) Lilium martagon L. Limodorum abortivum (L.) Lysimachia vulgaris L. Melissa officinalis L. Mellitis mellisophyllum L. Mercurialis annua L. Muscari botryoides (L.) Mycelis muralis (L.) Myosotis sparsiflora Mikan Neottia nidus-avis (L.) Nonnea pulla (L.) Ornithogalum montanum Cyr. Ornithogalum umbellatum L. + Peucedanum alsaticum L. Physospermum cornubiense (L.) Picris echioides L. Plantago media L. Polygonatum odoratum (Mill.) Potentilla alba L. Potentilla micranta Ramond. Potentilla neglecta Baumg. Primula veris L. Prunella vulgaris L. + Pteridium aquillinum (L.) Pulmonaria officinalis Pulmonaria rubra Schott Ranunculus auricomus L. Ranunculus fallax (Wimm. et Grab.) Ranunculus millefoliatus Vahl. Rumex crispus L. Rumex dentatus L. Rumex sanguineus L. Sanguisorba minor Scop.
20
U2
U3
U4
S1
S2
S4
R1
+
R2
+
R3
R4
+
+ + + +
+ + + + +
+ + +
+ + +
+
+
+ +
+ + +
+ +
+
+
+ + +
+ + +
+
+ +
+ + + +
+ + + +
+
+
+
+ +
+
+
+
+ +
+
+
+
+
+ + +
+
+
+
+ + + + +
+ + +
Ecology of the Sofia City CyanMagentaYellowBlack Page 21
The Sofia (Bulgaria) GLOBENET Sites: Description and Spatial Variation of the Landscape Mosaic Appendix I. Continued. U1 Sanicula europaea L. Silene alba (Mill.) Sonchus asper (L.) Stachys germanica L. Stachys recta L. Stellaria holostea L. Stellaria media (L.) Stellaria nemorum L. Symphytum tuberosum L. Tanacetum corymbosum (L.) Tanacetum vulgare L. Taraxacum officinale Web. Teucrium chamaedrys L. Thalictrum minus L. Torilis japonica (Houtt.) Trifolium medium L. Valeriana officinalis L. Veratrum lobelianum Benh. Verbascum phoeniceum L. Veronica chamaedrys L. Veronica serpyllifoli L. Vincetoxicum hirundinaria Medic. Viola reichenbachiana Jod. Viola riviniana Reichenb. Viola suavis M.B. Viola sieheana Becker Viscaria vulgaris Röhl.
U2
U3
U4
S1
+
S2
S4
+
R1
R2
R3
R4
+ +
+ + + +
+
+
+
+
+
+ + + +
+ + +
+
+
+ + +
+ +
+ + +
+ +
+ + +
+
+
+
+ + + +
21
Ecology of the Sofia City CyanMagentaYellowBlack Page 22
Lyubomir Penev et al. Appendix II A. Appearance of the Sofia GLOBENET sites (gradient 1) – Knyaz Borissovata Gradina (U1).
22
Ecology of the Sofia City CyanMagentaYellowBlack Page 23
The Sofia (Bulgaria) GLOBENET Sites: Description and Spatial Variation of the Landscape Mosaic Appendix II B. Appearance of the Sofia GLOBENET sites (gradient 1) – SE of German Village (S1).
23
Ecology of the Sofia City CyanMagentaYellowBlack Page 24
Lyubomir Penev et al. Appendix II C. Appearance of the Sofia GLOBENET sites (gradient 1) – S of Germanski Monastery (R1).
24
Ecology of the Sofia City CyanMagentaYellowBlack Page 25
Climate and Human Comfort of Sofia
Climate and Human Comfort of Sofia V. Andreev, Ch. Branzov, E. Koleva, A. Tsenkova, J. Ivantsheva, P. Videnov National Institute of Meteorology & Hydrology 1784 Sofia; Fax : (+3592) 988 44 94; E-mail:
[email protected]
SUMMARY Urban climate is a result of both the city’s geographical location and its structure, i.e. of regional meteorology, and its interaction with the fabrics. The continental background climate of the Sofia valley is described. Experimental investigations concerning the radiative inversion destruction in Sofia are also objects of consideration. The Sofia heat island effect is characterized by air temperature differences in urban and rural areas. Microclimatic regions of Sofia are also given. The lidar measurements of the spatial distribution of aerosol particles and the vertical profiles of temperature, humidity and wind are noted. They were used to investigate the diurnal evolution of the atmospheric boundary layer. Local wind deformations are investigated using laboratory and numerical models. Human comfort conditions in Sofia’s urban and suburban districts are given, analyzing the cooling index “K” distribution. As a result of a lack of correspondence with the climate peculiarities of the different parts of Sofia some defects of town-planning are noted. KEY WORDS Sofia city; Sofia valley; local meteorology; microclimate; urban climate; radiative temperature inversions; inversion destruction; urban heat island; Sofia microclimatic regions; lidar measurements; local wind field; laboratory modelling; numerical modelling; human comfort; cooling index.
INTRODUCTION The actual urban climate is strongly dependent on both the city geographical location and its structure, i.e. on the regional meteorology and its interaction with the human-made fabrics. That is why only a few micrometeorological rules are general enough to be exported from one city to another. L. Penev, J. Niemelä, D. J. Kotze & N. Chipev (Eds.) 2004. Ecology of the City of Sofia. Species and Communities in an Urban Environment, pp. 25-54 © PENSOFT Publishers Sofia-Moscow
25
Ecology of the Sofia City CyanMagentaYellowBlack Page 26
V. Andreev et al.
The most characteristic phenomenon of the urban climate is the urban heat island i.e. the systematic higher temperature of the atmosphere over the city compared to that of the upwind neighbouring rural area. It is known that the temperature differences DTu-r between typical urban and rural stations increase with time and are dependent on the city size, from 2°C for small towns up to 12°C for very large ones. Temperature is also quite variable during the diurnal cycle and dependent on the meteorological conditions with, for example, maximums at the end of the afternoon or the beginning of the night in conditions of weak winds and clear air, and in winter. A more detailed description of the recent knowledge about this urban climate peculiarity and, in particular, the physical processes that take part in the heat island generation can be found in the paper by V. Andreev in this book. Heat island parameters depend on the city size, structure, density, energy consumption and degree of pollution (see Mestayer & Anquetin, 1995; Oke, 1990; Oke et al., 1991). Where pollution is low, the impact of solar radiation on the urban surfaces seems to be the main factor. Where pollution is high, energy consumption is also usually quite high and the anthropogenic energy flux seems to be the determinant factor. Of course it is always a combination of various factors. At least 6 causes of heat build-up can be listed: the increase of long wave radiation from the sky; the decrease of solar radiation reflection due to radiative trapping; the increase of solar radiation absorption by urban materials due to their lower albedo; the increase of heat storage by urban wall and ground materials; the anthropogenic sources of heat and energy; and the decrease of latent heat consumption due to reduced vegetation and to the water collection networks. The role and importance of these various factors appear to be much dependent on the solar cycle and the season; e.g., the alteration of the water budget is a dominant factor in summer, while the dominant factor in winter seems to be the anthropogenic heat flux. Measurements over large urban areas show that the internal patterns of the heat islands are both rather complicated in space and changing quite rapidly in time during the diurnal cycle, as a result of the 3-dimensional influence of the terrain heterogeneity. It must be noted that while a negative temperature difference DTu-r is rarely observed in city centers, it happens quite often in suburban areas due to the presence of artificial ground surfaces and the absence of strong heat sources and heat storages (Myrup et al., 1993). This increases the actual complexity of the local flows over the city and the urban “ground” description for dispersion modeling. Observations show that the heat island appears to be a characteristic feature of the urban atmosphere and occurs over built-up areas of all sizes, even over small villages or isolated large districts, shopping centers etc. Experimental studies have demonstrated that the heat island effect is important not only close to the ground but also higher in the atmospheric boundary layer and that it generates either “heat bubbles” or “heat plumes” (Fig. 1). Due to the city-induced thermal and dynamic convection, the frequency and the intensity of the inversion layers tend to decrease over urban areas and the “urban boundary layer” therefore tends to be more often close to neutral. In conditions as in scheme Fig. 1a, the development of an “internal boundary layer” over a flat plate is characteristic due either to a change in roughness (dynamic internal layer) or to a change in the surface heat flux (thermal internal layer). These model flows have been extensively studied in wind tunnels. The scheme (a) sees the urban boundary layer as one (or a series of) internal layer(s) in development in the atmospheric boundary layer. Therefore the city constructions appear as a carpet of roughness 26
Ecology of the Sofia City CyanMagentaYellowBlack Page 27
Climate and Human Comfort of Sofia
elements and the city landscape as a rough terrain. Actually all meteorological conditions produce some combination of the two situations in Fig. 1 and the picture is much more complicated: the structure of the “urban boundary layer” is 3-D, depending largely on the local orography and on the heterogeneity of the different districts inside the city. In calm or weak wind conditions, the specific atmospheric circulation is formed over the city in complex terrain. Its characteristic components are “slope wind” and “urban breeze” (see review by V. Andreev in this book). The models of the boundary layer structure and air circulation over a city are classical but crude approximations. Most of the studies over the last 25-30 years however have been developed in the framework of these conceptual models. The urban ground is covered with natural and artificial materials, buildings and other obstacles. The heat budget as well as the flow structure close to the ground depend largely on the height and density of the constructions. The flow structure varies from a series of turbulent wakes due to isolated obstacles to strongly interacting turbulent streaks and even to turbulent flows having some similarities with flows in semi-porous media. The atmospheric layer below the roof level is called the urban sub-layer or urban canopy by analogy with the atmospheric layer between ground and the trees of dense forests. Kondo & Yamazawa (1986) proposed to define its height by the integral of the building heights: h=1/S(∑Hi.Si), where S is the total area under consideration, Hi and Si are the height and area of the individual obstacles (buildings). Urban areas can induce some meso-scale perturbations in the atmosphere. One of them is demonstrated by the urban “plume” model of Fig. 1a showing that the climatic perturbations generated by the city are transported by wind, like a mesoscale “plume” of anthropogenicallymodified and polluted air as large as the city itself. This plume can generate a series of perturbations over the downwind regions at distances of several hundreds of kilometers (e.g., the “Ozone plumes”during photo-oxydant episodes). a)
V > 3 m/s Urban “plume” model Urban boundary layer
Urban canopy or sub-layer
Rural b)
Suburbs
City center
Suburbs
Rural
V < 3 m/s Urban “dome” model Urban dome
Rural
Suburbs
City center
Suburbs
Rural
Fig. 1. Models of the urban heat island in conditions of (a) moderate to strong wind and (b) weak wind.
27
Ecology of the Sofia City CyanMagentaYellowBlack Page 28
V. Andreev et al.
Humans are constantly being exposed to the simultaneous influence of meteorological (temperature, air humidity, wind speed, etc.) and geophysical (solar radiation, Earth magnetism) impacts. In addition, their mutual action is not equivalent to the sum of their separate actions. There is a complex connection between every element of urbanization (a building, a street, a plant, a park, a parking lot, etc.) and the environment in which it exists. As a result, each element or a complex of elements form their own, specific microclimate. The alteration of one of the basic abiotic factors of the city ecosystem - the climate, because of city development, leads to a change in the degree of comfortableness of the climatic conditions for the city population. A range of indices and bioclimatic parameters have been introduced through which attempts have been made for the assessment of the impact of the climatic conditions on humans. Most essential for humans are thermal conditions. Their effects are revealed in the process of heat and mass exchange with the environment. Humans represent a complex thermodynamic system, supporting a high stability of its inner isotherm, in spite of substantial variations of external conditions. During conditions of comfort, the heat balance is not disturbed, the energetic balance of a human is reached most easily and the vitality of the human being is at a normal level. In other words, heat comfort is present at a combination of the meteorological conditions when the thermoregulation of the organism is minimally burdened, i.e. the organism is at physiological rest. There are different approaches and a multitude of methods, indices and complex indicators that have been developed for rendering an account of the influence of weather on humans. The conditions of comfortableness for the human organism are evaluated by using these methods or a bioclimatic regionalization of a certain territory is performed. The oldest method used for a complex assessment of the thermal conditions of the environment is the method of effective temperatures. Effective temperature, suggested by Houghten & Yaglou (1923), represents the temperature of static air with saturated water vapour, in which a human being experiences the subjectively same sense of comfort as in the real environment for which the effective temperature is calculated. Another similar parameter is the effectively-equivalent temperature. It is determined by the simultaneously measured values of the basic meteorological elements – temperature, air humidity, and wind speed. Analysing the existing indices for assessment of the climate comfort conditions during summer months, Davis (1968) suggested an index based on data for the average diurnal maximum temperature (in deg F) for the three summer months (June, July, August), on the average diurnal sun radiation (in hours for these months) and the total precipitation (in inches). Attempts have been made for the completion of maps of physiological comfort. Gregorczuk & Cena (1967) calculated the effective temperatures for the territory of the entire Earth for January and July. Maunder in 1962 (see Temnikova, 1975) used thirteen climate indicators for determining the favourable and unfavourable conditions for living regions. Terjung (1966, 1968) and Terjung & Louie 1971 introduced the term “index of comfort” for the USA, later applied to the entire Earth. The combination of temperature and relative air humidity form its foundation. 28
Ecology of the Sofia City CyanMagentaYellowBlack Page 29
Climate and Human Comfort of Sofia
B.A. Eisenstaedt in 1973 (see Temnikova, 1975) developed a method taking into account the body heat balance through the quantity of released perspiration per a unit time, i.e. through an objective criterion for reflecting the degree of tension of thermo-regulation. The method of I.A. Arnoldi (see Temnikova, 1975) can be applied for determining the cooling action of the wind by the so-called “conditional temperature”. It is accepted by the method that an increase of wind speed by 1m/s is equivalent to a decrease in air temperature of 2°C. For a description of the heat balance of the human body, Rusanov (1973) united the various existing meteorological complexes into classes of weather, composing different types of weather in relation to their influence on the human heat state. Evaporation, especially perspiration, is a strong body cooling process and can be determined by the temperature-humidity index THI of Thom (1959). There exists a method classifying weather in relation to its effect on the psychic state of man (Besancenot, 1978). This method, based on the duration of sun radiation and the quantity of precipitation, evaluates the dominant climatic influences on the human psyche. The climatic-physiological approach, suggested by Kandor et al. (1974) is applicable for the examination of the climatic influence on the human organism for the purposes of summer recreation and tourism. In the method of vapour pressure of Marinov (1971), an important role is played by the so-called normal physiological boundary of heat adaptation. By means of this method, the following energetic states of the organism are determined: cooling – the pressure of water vapour on the body surface is equal to air pressure; subcomfortable – the pressure of water vapour on the body surface is between air pressure and 31 hPa; comfortable – the pressure of the water vapour on the body surface is between 32 and 44 hPa; over-comfortable – the pressure of the water vapour on the surface of the body is from 45 to 62.5 hPa; overheating – the pressure of the water vapour is over 62.5 hPa. The conditions of comfortableness in the urban environment, situated in a complex relief with a different degree of urbanization, changing with the development of the city, should be characterised by a method based on these meteorological elements which are greatly influenced by urbanization and relief. For some regions, as, for example, Sofia, the most appropriate index appears to be “the coefficient of cooling” K, suggested and used by Siple & Passel (1945) (see also Kyle, 1994): K=1.163(10.45+10÷v-v)(33-t), [W/m2] (1) where v is wind speed in m/s, t is the air temperature in deg C. This index characterizes the degree of body cooling due to the influence of wind at a certain air temperature in the shade and average skin temperature of 33°C, without taking into account evaporation. The sizeable changes in local, urban and regional climates due to anthropogenic activities have produced many changes in the biosphere and a mixture of winners and losers in the socioeconomic context (see Mestayer & Anquetin, 1995). In the lee side of St. Louis, where the climate had been altered, 25% increase in average summer precipitation and a reduction in visibility due to pollutants have led to more automobile accidents, less leisure time, increased storm damages, more suburban and farm land flooding, and more soil erosion, but a net increase in crop production and increased cropland values. The migration away from urban centers leading to suburban development over the last 100 years was partly due to the less-than-pleasant climate, including 29
Ecology of the Sofia City CyanMagentaYellowBlack Page 30
V. Andreev et al.
poor air quality and worse human comfort in the urban center. Study of the relocation process of wealthier urban residents and the stay-in-place decisions of affected farmers located beyond the cities is important. These findings are important for those concerned with the environmental impacts and socio-economic adjustments to large-scale changed climatic conditions.
MATERIAL AND METHODS Our present knowledge of urban climatology is based mainly on three types of data and methods: site measurements; wind tunnel and water tank simulations and numerical model studies. It must be noted that a large part of our knowledge comes from field campaigns. Useful measurements of the urban atmospheric phenomena that can add to these three types of data are the measurements by lidars, sodars, radar profilers and satellites. For characterizing the general climatic conditions and the comfortableness of the air environment in the city of Sofia and the Sofia valley region, perennial data, received from climatic stations in the region performing measurements at 7, 14 and 21 hrs local time, was used. The microclimatic studies are substantially more expensive for they require a climatic network with an increased density. In Sofia, a similar network with a comparatively good density exists. Nevertheless, the number of stations of this network decreased considerably in the last 10-15 years due to shortage in financing. This fact substantially limits the possibilities for scientific studies and especially for those concerned with the microclimate of the city. For some of the studies, as, for example, for receiving the climatic characteristics of the temperature inversions in the Sofia valley, data from 1953-1963 from aerological sounding and hourly data from the thermographs, situated at 2 and 40 m (on the meteorological tower of NIMH - in Mladost 1), as well as synoptic maps for a 5-year period were used (see Blaskova et al. 1968). The high price of regular microclimatic observations necessitates the implementation of other types of observations and most of all of short-term “field (natural) experiments”, frequently called “field campaigns”. For these, temporary networks of an increased number of stations over short periods of time, but with more frequent observations, are used. Due to the substantial influence of microclimatic conditions on air quality in local regions, these experiments are usually complex, i.e. the meteorological measurements (in surface layer and in height) are combined with measurements of the concentrations of air pollutants (see Palmgren, 1996). The experiments are usually conducted during the representative periods (from 15 days to one or more months) of the cold and warm half-year. The number of stations in the examined region as well as the frequency and variety of measurements are increased. In this publication, altogether with results from studies from perennial regular climatic observations, data obtained from short-term experiments is also presented. Some of these observations were implemented through the entire city (e.g. “Sofia’76” – see Andreev, 1980 or “Sofia’82” – see Andreev, 1984) while in others – only in separate regions (Sofia-Drujba’93 – see Syrakov, 1992, 1993), and still in others – only at separate points (see Branzov et al., 1992). The last type – an experiment at a discrete point – was conducted in 1988-1991 in the NIMH region (Mladost 1) within the project “Experimental study of the structure and dynamics of temperature inversions and the convective boundary layer for the Sofia 30
Ecology of the Sofia City CyanMagentaYellowBlack Page 31
Climate and Human Comfort of Sofia
region”. The project was performed in two observational periods per year – spring and autumn, during around 20 selected days with ground radiative temperature inversion. The measurements began before sunrise, when the inversion reaches maximum thickness, and continued to its destruction in the examined layer. Thus, for the separate experimental days, from 7 to 12 vertical profiles of temperature, relative air humidity and wind speed were obtained. For the layer from 2 to 600 m, pilot-balloon and a specially constructed radiosonde were used (Branzov et al., 1992). The measurements were made every 50 m as the entire measurement with the observation balloon in the 2-600 m layer continued for 40 min on average. That is why interpolative methods were used for measuring the values of temperature, humidity and wind speed at the same moment in time. For the 600-1500 m layer, data from the routine NIMH soundings in 2:00, 8:00 and 14:00 hrs local time were used. Computer experiments with numerical models were used to replace the absence of direct observations in the natural atmosphere. The oldest model simulations of the perturbations induced by large cities on the meso-scale climatology were realized with meso-scale models. The stationary versions of these models did not have the features allowing to simulate or predict the climatology of the cities in detail. But, the recent development of non-hydrostatic meso-scale models allowed us to perform such studies. Such experiments for the Sofia region were made by quasi-hydrostatic meso-scale model developed by Ganev et al. (see Ganev, 1993). This model was adapted for the Sofia valley (Dimitrova & Ganev, 2000) and used for determining the meteorological fields, having a significant role for the meso-scale distribution of air pollution in the valley at stationary conditions.
RESULTS AND DISCUSSION The city of Sofia is situated in the central and southern parts of the Sofia hollow valley at a mean altitude of about 550 m. The Sofia hollow valley itself consists of two parts: hollow’s bottom (often called “field”) and the surrounding mountain slopes. Climate in this region of Bulgaria has a continental character. What are the most characteristic features of the climate of the Sofia valley? The quantity of radiative energy reaching the ground surface depends on many factors – the geographical latitude and ground exposition, the cloud cover regime, composition of air and in particular the aerosol particles. The main part of the income of radiative energy on the ground surface is due to direct solar radiation. The quantity and the type of its diurnal and annual changes depend mostly on the height of the sun. The maximum is during noon. The intensity of direct solar radiation on a horizontal surface in the region of Sofia at noon changes from 0.24 kW/m2 in winter to 0.70 kW/m2 in summer. The total solar radiation has a well expressed annual change. Its annual value on average is 136 MJ/m2. These values are lowest in December-January – around 3.5-4.5 MJ/ m2, while in a sunny January they can reach 6-7 MJ/m2. Because of the proximity of the mountain, the sun sets in winter 1-2 hours earlier than in other seasons and this influences incoming solar radiation. Solar radiation reaches its maximum in July when on average it is 19.3 MJ/m2. In separate years it can reach up to 27 MJ/m2. Spring is sunnier than autumn; in April the solar radiation is around 14 MJ/m2, while in October it is 9 MJ/m2. Local factors greatly influence the values of solar 31
Ecology of the Sofia City CyanMagentaYellowBlack Page 32
V. Andreev et al.
radiation. As an example, solar radiation on many steep, southern slopes is almost always at a maximum. A tendency exists for a decrease in the total incoming solar radiation during the 1950-1990 period. In the last 10 years an increase of this radiation is observed (Fig. 2). The duration of sunshine is different through the seasons and depends on the length of the day and the cloud cover regime. The average duration is 2040 hours annually. In separate years
MJ/m2
200
Sofia
150
100 1950
1960
1970
1980
1990
2000
year Fig. 2. Change of annual total radiation.
Annual sun shine duration
2600
hour
2400 2200 2000 1800 1950
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
year Fig. 3. Change of annual duration of radiation. 80
Sofia
mm
60 40 20 0 I
II
III
IV
V
month Fig. 4. Annual distribution of maximal precipitation.
32
VI
VII average
VIII abs. Max
IX
X
Ecology of the Sofia City CyanMagentaYellowBlack Page 33
Climate and Human Comfort of Sofia
the annual duration reaches 2600 hours. It is sunniest in July and August when the duration of sunshine is 280-300 hours and it may reach up to 360 hours monthly. The duration is shortest in December and January – around 75-90 hours and in separate years as low as 25-35 hours. On Fig. 3 a well expressed tendency of increasing yearly sunshine duration over the last years can be seen, as well as for total solar radiation. Annual change in precipitation in this part of the country has a well expressed continental character – the quantity of winter precipitation is substantially lower than in summer (Fig. 4); the difference is around 16% from the annual total precipitation. Winter precipitation in the region is lowest – 100-120 mm (or 18% on average of the annual total precipitation). Winter precipitation decreases from December to February (in February – 30-40 mm), but the difference between the two months does not exceed 10-15 mm. In the winter months, though rarely, greater precipitation may occurs, reaching 60-70 mm in 24 hours. In winter, around 50-60% of the precipitation is from snow. In spring, precipitation increases in comparison with winter. In March it is still low and close to that in February due to the humidity-scarce continental air masses. At the beginning of spring snowy precipitation prevail. In the second half of spring and in summer the transport of air mass from the Atlantic ocean is more frequent, appearing as cold fronts. The Atlantic air is rich in moisture and its intense warming over the fastly heating mainland leads to a strong increase in instability. That is why there is generally shower-type precipitation, frequently accompanied by thunderstorms. Frontless-type precipitation is also more frequent. The increase in precipitation is especially greater on the northern slopes of the surrounding mountains due to the strengthening of the vertical component of the air flux from forced lifting. The increase in precipitation in April (still a part of it is snow) is relatively low, but in May it increases sharply by 2530 mm, reaching 70-90 mm. The summer precipitation is generally frontless and the monthly precipitation sums are greatest – around 170-190 mm (28-30% of the annual total precipitation) with highest values in June (75-85 mm). A gradual transition to winter circulation occurs in the autumn. The influxes of Atlantic air, decreasing already in August, continue to diminish in September. The sum of precipitation in the region in September is lower than in August but is still relatively high (35-75 mm). In October and especially in November, the cyclonic activity in the Mediterranian basin is intensified and the periods with lasting and heavy precipitation in the country are more frequent. The autumn increase of precipitation is not well expressed in the region. More, precipitation in October is almost equal to that in September and on locations even lower. In November precipitation is 1020 mm lower than in September, and 40-60% of it consists of snow. Monthly precipitation varies widely in different years and ranges between 0 and 100 mm, and sometimes more. The frequency of a given quantity of precipitation in separate months is different. In January the probability for the monthly sum to be below 20 mm is around 20%, while in May and in June – only 5%. In May-June the probability is greatest for sums over 100 mm; it is 3040%. The number of days with precipitation has a weakly expressed annual change. In January there are around 10 days with precipitation, rising to 15 days in May and June and then it decreases again. On average there are 120 days with precipitation within the year – this number has decreased slightly in recent years. Over a 24 hour period, different quantities of precipitation are observed. Of the 120 days with precipitation, for 5-6 days it is over 15 mm, and in only 2-3 days it exceeds 25 mm. In separate 33
Ecology of the Sofia City CyanMagentaYellowBlack Page 34
V. Andreev et al.
years the diurnal precipitation can surpass the monthly norm. Greater diurnal precipitation falls during the warm parts of the year because it is mostly of the convective type. The diurnal maximum of precipitation and particularly its extremes are generally the result of accidental processes, frequently of local character. Climate dryness is determined by the distribution of meteorological elements, in particular temperature and precipitation. To follow the perennial change of precipitation in the region, the average precipitation for this region is usually calculated, as well as an index of anomaly (Koleva, 1988):
Vj =
1 n xi ∑ n i =1 xi
where j = 1,..., N; xi – the annual precipitation at the ith-station, xi – the average annual precipitation for the same station, and n – the number of stations. In Fig. 5, the change of the annual index of anomaly is given, calculated at 9 stations. The chart shows a tendency for a decrease in precipitation in recent years. The last few years in Bulgaria have been drier and warmer than normal (Koleva et al., 1996). According to the World Meteorological Organisation, the averages of the 1961-1990 period are used to describe the contemporary climate. Precipitation is around 80-90% bellow normal. Winters and summers are warmer than usual. The perennial change of precipitation sub-periods can be defined, characterised by different precipitation conditions. During the last hundred years, there were two periods at longer-lasting and strong droughts – 1942-1953 and 1982-1994. During the first period, the dry years are 26% while in the second they increased to 32%. The dry years for the 1961-1990 period are only 10%. The average total annual precipitation for the 1961-1990 period is 543 mm, while for the 19821994 – 473 mm. Analysis of the annual distribution of precipitation during these periods shows that during the 1942-1953 period drought was observed at the end of summer and the beginning of autumn when precipitation was 10 mm bellow normal. In the 1982-1994 period, the precipitation is particularly low also in winter. In all three periods, precipitation maintained its moderatecontinental distribution.
Anomaly index
1.6
Sofia plain
1.2
0.8
0.4 1930
1940
1950
1960
1970
1980
1990
2000
year Fig. 5. Change in the annual anomally index, linear trend and approximation with a 5-degree polynomial.
34
Ecology of the Sofia City CyanMagentaYellowBlack Page 35
Climate and Human Comfort of Sofia
The driest year during the 20th century for the Sofia valley is 2000. Among the driest years are also 1945, 1934, 1946, 1990, 1993, 1994, 1985. After each dry period, there usually follows a comparatively moist period. During the last decade, 1991 and 1995 were substantially humid. For determining the frequency and intensity of droughts, different criteria are used such as the following: , where DT and DP are the deviations of the air temperature
a) Ped Index:
and the precipitation from their average values for a given period of time; sT , sP – the standart deviations of these elements. Values for the index between 1 and 2 show a weak drought, at 2
3 the drought is severe. Negative values for the index point to a moist period. During the last decade, the dry years were 1989, 1990, 1993, 1994, 2000. The years 1990, 1994 & 2000 were considerably dry, and the values for the dryness index were greater than 4. During the last years, a tendency exists for an increase in the Ped Index (Fig. 6). b) De Marton index: J =
12P , where P and T are the monthly precipitation and air temT + 10
perature. At values for the index J lower than 30, there is a drought, and at values below 20 – a severe drought. The average annual value of this index in the region of Sofia for the 1961-1990 was 28-29, while during 1982-1994 it decreased to 23-25. On Fig. 7, the change of the Marton index is given per month. The driest months for 1982-1994 were July-October, when the index J reached values between 2 and 17. ∆T ∆P Ped = − Air temperature for the region increased over time. This tendency is well expressed both for σ T the σ Paverage temperature and the minimal and maximal temperatures (Fig. 8 and Fig. 9). The same is observed with other temperature characteristics. Thus, for example, the number of days with a maximal temperature above 30°C also grows in the recent years (Fig. 8). Temperature inversions are the background of a range of phenomena and processes such as the convective boundary layer (CBL); gravitational waves; accumulation of polluting substances in the atmosphere; minimal night temperatures accompanied by ground frost, etc. The mutual connection and interdependence among these phenomena as well as the role of the ground surface is a matter of interest for every region. Climatic characteristics of the temperature inversions in a particular region are of independent interest in connection with some
Red index
6
Sofia
3 0 -3 -6 1900
1910
1920
1930
1940
1950
1960
1970
1980
1990
2000
year Fig. 6. Perennial change of the Red index.
35
Ecology of the Sofia City CyanMagentaYellowBlack Page 36
Marton index
V. Andreev et al. 70 50 30 10
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
XII
month 61-90
42-53
82-94
number of days
Fig. 7. Annual distribution of the de Marton index for different periods.
Sofia
40 30 20 10 0 1950
1960
1970
1980
1990
2000
year
temperature °C
Fig. 8. Perennial change in number of days with maximum temperatures over 30°C and smoothing with a 5-degree polynomial. 7 6 5 4 3 1950
Sofia
1960
1970
1980
1990
2000
year Fig. 9. Perennial change of the average minimum temperature and linear trend.
tasks and particularly for the control of air pollution. Such characteristics for the Sofia valley were obtained by Bluskova et al. (1968) from 10-years’ observations of the ground air (at 2 and 40 m above ground level) and by aerological observations. It was found that around 90% of all inversions are situated at heights up to 2500 m. Within a year, the greatest frequency (around 45% of all inversions) is for those starting from the ground surface or the first 100 m; around 30% of the inversions are in the 200 to 1400 m layer, and 25% – in the layers above 1400 m. The most frequent ground inversions are those having a thickness of 200-400 m (35%), and with thickness over 1000 m – only 8%. The average duration of the inversions reaching 40 m in height is around 19-20 hours in winter and 10-11 hours in summer. The average duration is greatest in January and December – around 23 hours for inversions up to 40m and around 25 hours for up to 1400 m. On separate occasions, the duration of continuous inversions may 36
Ecology of the Sofia City CyanMagentaYellowBlack Page 37
Climate and Human Comfort of Sofia
reach 10-12 days in winter. Inversions in the ground layer (40 m) in winter lasting over 6 hours were present in 80% of all days in a year, while in the layer up to 1400 m with the same duration – in around 60% of all days. The number of positive air temperature gradients in the layer up to 300 m was almost 7 times greater than the negative ones and the most frequently measured values were 0.6-0.8 deg/100 m. With an increase in altitude (in the 300-600 m and 600-900 m layers) the most frequently measured gradients were 0.5-0.6 deg/100 m. Inversions in the layer up to 40m for all months of the year, are formed during the period beginning from 10° Sun’s altitude to sunset. The destruction of the inversions in this layer happens usually when the Sun rises to 15°. Inversions are formed at all forms of the surface pressure field but most frequently at anticyclonic fields (about 60% of all days). At cyclonic field conditions they are present in around 30%, and at transitional pressure fields – in around 10% of all days. In the publication by Bluskova et al. (1968), research was conducted for the disease rate based on data for the districts Hadji Dimitar and Dragalevci at the presence of inversions when an increase of air pollution is usually observed. It was found that, for example, the number of sick children and adults with rhinopharyngitis, tonsilitis and bronchitis increases on average by 2030% in winter. The study of Teneva et al. (1989) complemented our knowledge of the temperature inversions in the region of Sofia, reporting some supplementary characteristics separately for each of three radiosoundings at 02, 08 and 14 hrs local time for the 1970-1979 period. Thus, for instance, ground inversions according to the 02 hrs sounding have their greatest frequency in August (around 56%) and lowest - in February (around 34%). At 08 hrs, though, these inversions are most frequent in December (around 38%) and least frequent in July (only 3%). The ground inversions at 14 hrs during all months are rare (frequency bellow 1%) with the exclusion of December when the frequency is around 12%. The average thickness of the ground inversions is correspondingly: 200-300 m at 02 hrs, 320-450 m at 08 hrs, while at 14 hrs there are separate events with thicknesses from 100 to 480 m. Their mean intesity varies from 1.7 to 2.6°C at 02 hrs and from 0.7 to 3.1°C at 08 hrs. The ground inversions in Sofia at 02 and 08 hrs during the whole year are most frequently observed during calm weather at ground layer (61%). Their frequency is not that low at a weak southern wind (around 19.5%) which is explained by mountain-valley wind from the Vitosha mountains at small-gradient pressure fields. The upper inversions (with a lower boundary of 2000 m) at 02 hrs are most frequent in February (around 43%) and less frequent in August (around 14.5%). Their frequency increases at 08 hrs with a maximum in summer months due to the elevation of part of the ground inversions observed at 02 hrs. The upper inversions are a very frequent phenomenon at 14 hrs in autumn and winter (in around 76% of all days with inversions). Their average thickness is between 290 and 520m and their average intensity reaches up to 4.3°C with a maximum in winter. The altitude of the lower boundary of the upper inversions has great importance for ground air pollution. The inversions with a lower boundary below 1 km at 02 hrs are around 73% of all upper inversions. They are observed most frequently in winter (32% at 02 hrs, 30.5% at 08 hrs, and 38.5% at 14 hrs), and are less frequent in summer (13.5% at 02 hrs, 4% at 08 hrs) and spring (around 27% at 14 hrs). Some tasks, connected mainly with the diffusion and transformation of admixtures in air as, for example, the prognosis for the level of air pollution in a certain region, demand knowledge 37
Ecology of the Sofia City CyanMagentaYellowBlack Page 38
V. Andreev et al.
concerning the structure and mechanisms of destruction of the radiative temperature inversions. During the 1988-1991 period, a project was realized for studies in this field, named “Experimental study of the structure and dynamics of temperature inversions and the convective boundary layer for the Sofia region” (see Branzov et al., 1992). In Fig. 10, an example of the transformation with time of a radiative temperature inversion at the time of that project is shown. The vertical profile of temperature is in hours towards sunrise (time SR). The measurements were made with an observation balloon and a special meteorological sounding device. Because the time for a vertical section is around 30 min., the momental expression of the vertical temperature profiles in Fig. 10 is received with an interpolation on time with a cubic spline. The project shows (see Branzov & Panchev, 1991) that a ground radiative temperature inversion in Sofia is not observed at wind speeds over 3 m/s at 10 m. The altitude of the upper boundary of the ground radiative inversion, at unchanged conditions, becomes greater with every following day of the period, when it exists incessantly. Also the speed of formation of the layer with temperature inversion in autumn is greater than in spring. The dependence of altitude of the upper boundary of the temperature inversion on the sequence of the day in the series of days with temperature inversions, was used for the creation of a semi-empirical model for determining the altitude of the unstable convective boundary layer and the time for destruction of a ground radiative inversion over Sofia. It is applicable to wind speeds up to 6 m/s. The physical model for the development of a ground radiative temperature inversion (Branzov et al., 1991) suggests: radiation transfer of heat, determined by the thermo-conductivity equation in a onedimensional case, at a constant thermo-conductivity coefficient; zero heat flux of the upper boundary level of the temperature inversion; and a linear profile of temperature at the initial Vertical profile of temperature in Sofia, 26.10.1989
450 400
Height [m]
350 time SR-1 time SR+0 time SR+1 time SR+2 time SR+3 time SR+4 time SR+5 time SR+6
300 250 200 150 100 50 0
5
7.5
10
12.5
15
17.5
20
Temperature [deg C] Fig. 10. Experimental results from the transformation of the vertical temperature profile depending on timing after sunrise in a day with a temperature inversion.
38
Ecology of the Sofia City CyanMagentaYellowBlack Page 39
Climate and Human Comfort of Sofia
moment. Input parameters are: thermo-conductivity coefficient, the surface air temperature (at a height of 2 m) at the initial moment and in separate moments of the inversion development and for the altitude of its upper boundary. The input data for determining the night development of a ground radiative inversion are the initial moment of development; the temperature at 2 m at the initial moment and after that at random time intervals; the time of sunrise and the temperature at 2 m at the moment of sunrise. It is interesting to note that anticyclonic weather occurs during two thirds of days in the year. In this type of weather, the microclimatic differences, as a result of the relief ’s peculiarities and city complex influence, can be most clearly distinguished. Hollow forms of the terrain, the relatively high altitude above sea-level and city complexes enhance continental features of the city’s climate. Let us now review in short the changes of the basic climateforming factors and the characteristics of the background continental climate due to these peculiarities of the region, i.e. to give a notion for the microclimate over Sofia. This was done from data of the basic meteorological elements from the functioning (until a few years ago) climatic stations in Sofia and the Sofia valley, as well as from the complex experiments performed in the region. The heat island is most frequently characterized by a temperature difference “city-surroundings” or “urban-rural” stations – ∆Τu-r. Blaskova et al. (1983) found, from data from climatic stations functioning temporarily for 12 years (1954-1965), that the average annual temperature in Sofia is 0.7-0.8°C higher than outside the city. The differences are more substantial in the average monthly temperatures for observation at 21 hrs when it is 0.5-1.4°C. Zlatkova and Tzenkova (see Andreev, 1980 or Blaskova et al., 1983) found from expedition data in June and December 1976 that the heat island in summer is most intensive during 03-04 hrs at night when ∆Τu-r = 6.3°C was registered. It fades away during the hours around the temperature maximum-14 hrs. In winter the greatest differences on average are found at 22 hrs. The maximal intensity of the heat island registered during the winter experiment was ∆Τu-r = 4.3°C. In the studies by Syrakova and Zaharieva (1998, 1999) the diurnal and annual changes of ∆Τ for a ten-year period (1981-1990) were studied on hourly data from the stations in Sofia the Botanic garden (Levski sq. – urban station), NIMH (suburban station) and Lozen (rural station). It was found that an intensive heat island appeared most frequently at night in the summer period July-September where in more than 20% of time the average value of the difference among the three stations ∆Τ exceeded 3°C, reaching 3.2°C. The largest values of ∆Τ reach almost 10°C. In some occasions not only a heat island is observed but, contrary, a “cold island” (∆Τ < 0°). The average frequency of a negative temperature difference is 20% most frequently appearing in July-September. The local weather conditions (wind speed and cloud cover) greatly influence the urban heat island. When there are fewer clouds and a low wind, ∆Τu-r in summer nights reach values greater than 4oC. An increase in wind speeds leads to a significant decreasing of the heat island intensity, and combined with an increase in cloud cover results in an almost complete loss of the temporal pattern features. With the expedition data it is possible to determine the form and structure of the heat island for Sofia. It is prolonged in the N-NE sector in direct dependence of the positioning of industrial sites and the “Kremikovtzi” plants. The main nucleus, i.e. the area of most intensive heating, is 39
Ecology of the Sofia City CyanMagentaYellowBlack Page 40
V. Andreev et al.
observed over the central city parts that are densely built-up, with intensive public transport and high dust concentrations. For Sofia, a second nucleus is characteristic over the Iskar railway station – a city district with industrial sites, living areas and a recreational zone (Fig. 11). ∆Τ only gives a general view of the intensity of the heat island, but not of its structure. In most cities, with a million people there exists a constantly functioning network of meteorological observation sites and from these data one can obtain adequate information on the microclimatic conditions in a city. In Sofia, unfortunately, most meteorological stations were closed, including the first Bulgarian station (at the Levski monument). The field experiments conducted in 1992 and 1993 give an idea of the thermal regime in the eastern part of the city (Batchvarova et al., 1996). The intensity and configuration of the zones with higher temperatures (here the heat island is not mentioned because the research is conducted only in the eastern part of the city, where a second heat nucleus is formed) depend on the meteorological conditions. The intensity of these zones is greater at night and it decreases two fold at noon. The horizontal gradients are maximal at the boundary between urban and rural environments, and decrease within the urbanized territory. The maximal registered difference “city-surroundings” ∆Τu-r is 4.8°C during summer and 4.0°C during winter observation periods.
district Iliqnci district Obelia
district Benkovski
district Vrazdebna
district Suhodol - 1. 0
-2 .0
district Knjazevo
district Bojana
-3 .0 -4 .0
Gara Iskar district Darvenica
village Gorubljane district Dragalevci
Fig. 11. Air temperature distribution in Sofia on the morning (8 h a.m.) of 30.11.1976.
40
Ecology of the Sofia City CyanMagentaYellowBlack Page 41
Climate and Human Comfort of Sofia
The wind regime in Sofia was mainly studied from data received through complex experiments and mobile measurements. According to Zlatkova and Tzenkova (see Tzenkova & Ivancheva, 1999), wind speed in different regions of the city is, on average, 1.3-2.0 m/s lower than outside the city. This difference has a maximum during afternoon hours (14-16 hrs) and at a minimum in the morning. As a quantitative indicator for the change of wind in different parts of the city, the change in wind coefficient was used – Kv = Vi/V0, where Vi is wind speed at a random city point and V0 – the wind speed in the selected rural point. The change in wind coefficient provides a possibility to choose such building forms allowing for the creation of a more favorable wind regime in a certain location. Zones, where Kv is from 0.0 to 0.5, are considered having weak ventilation. In places, where Kv is from 0.6 to 1.0, there is optimal ventilation for the city environment, and where Kv>1.0, the region is characterized by strong winds. The mobile measurements conducted in 1983, 1987 and 1989 for the distribution of wind speed for Sofia show, that in most parts of the city Kv<1, as there are a number of locations where Kv=0. These results confirm the decreasing influence of the city on wind speed. In conditions of anticyclonic weather in summer, for the southern districts of Sofia, situated at the foot of the northern slope of Vitosha, the appearance of “slope wind”, also called “mountain breeze”, is characteristic. During daytime, close to the earth surface, it should blow from the city upwards on the mountain slope (anabatic wind), but this event is almost non-occurent. The night slope wind, which is a katabatic type wind, is well expressed especially in July and August, beginning after 19 hrs and calming down around 07 hrs. During calm and clearly anticyclonic weather at night, this wind brings clean air, rich in negative ions, from the mountain slope towards the polluted city parts. Because its vertical growth (up to 30-40 m) and speed (1-3 m/s) are small, the high buildings situated crosswise can stop the clean air at the periphery of the city. A sound example of this is the mobile measurements conducted on 22 July 1987. The values of Kv obtained by these measurements of wind speed in different city points, showed that while at 20 hrs in the region of Dragalevci the slope wind was clearly expressed with speeds over 2 m/s and Kv=1.2, in the Hladilnika and Krusta region Kv was already bellow 1. This appeared even though the points were situated in close proximity to a boulevard, coinciding with the direction of the wind, and only a few meters away in the side-streets there was no wind. The same experiment illustrated the role of the mountain, the green zones and the aeration conditions for the thermal regime. From 35.4°C at the National Palace of Culture, the temperature gradually fell to 35°C at Krusta, 34.3°C at the ring-road to 29.6°C over Dragalevci, while in the center of the city at the crossroad of Ekzarch Josiph str. and Rakovska str., it maintained a value of around 38.5°C. Western and northwestern winds usually prevail in the city of Sofia. Therefore, in 1988 and 1989 mobile measurements of the distribution of the wind flux were conducted in some parts of the city during a northwestern wind. Inside the relatively densely urbanized windward part of the city (the Vitosha municipality) a process of blowing-through was observed, and in open areas (the National Library and the Levski monument) the wind maintained its speed. In the southeastern periphery of the city (the background NIMH station in Mladost1 district), the wind completely died away. The situation is rather different in streets having an orientation close to that of the wind flux. Here, the wind main41
Ecology of the Sofia City CyanMagentaYellowBlack Page 42
V. Andreev et al.
tained its speed or, in some cases as on the ring-road, an effect close to tunneling appeared and the wind speeds increased. Relative humidity in the densely urbanized center of the city at the time of all field studies in Sofia is lower than in the suburban areas. According to data from experimental researches in the summer and winter of 1976, the difference (Df) in relative humidity between Vazrazdane sq. and different observation points at the periphery of the city has an almost identical diurnal variation. Differences were greatest in the hours before noon and after sunset, when Df reached up to 15% in both seasons. The mobile measurements in October 1989 showed, that, on separate occasions the difference in relative humidity between locations within the city, characterized by different ground surfaces, can reach up to 50%. Microclimatic regions of Sofia. The relief ’s peculiarities and city complex influence enhanced continental features of the city’s climate, and the mountain surroundings create possibilities for fast spatial changes (basically horizontal gradients) of both the meteorological elements and air pollution characteristics. If the terrain is entirely flat, such changes occur at much greater distances. The indicated factors lead to the formation of Sofia City’s specific microclimate. In the presence of such peculiarities, the City can be divided into both, microclimatic regions and air pollution levels. According to studies carried out by the National Institute of Meteorology and Hydrology (NIMH) (Blaskova et al., 1983), there are four microclimatic regions in the city of Sofia and the Sofia hollow valley: 1) open hollow field; 2) central city part; 3) peripheral part of the city; 4) foot of the mountains. The first microclimatic region “Open hollow field” is the lowest part of the Sofia valley, at 520-560 m altitude. It is surrounded by mid-altitude (Vitosha and Stara planina mountains) and low-altitude mountains (Ljulin, Viskjar, Lozenska and Vakarelska). The local climatic conditions in this region are frequently unfavourable for the dispersing of pollutants in air. A reason for this is the frequent and long-lasting temperature inversions. Thus, for example, in 75% of the days in winter in the layer up to 2000 m altitude, there is some inversion with a duration of up to 3-4 24-hours periods. The average thickness of the ground inversions in winter is 350-400 m. In these cases the conditions are unfavourable for the dispersion of pollutants both from low and high sources, because the smoke plume from even the highest chimneys remains in the inversion. The temperature inversions are a frequent phenomenon also in summer, but then they are less thick and of shorter duration – they are destroyed in the morning hours. They also obstruct the dispersion of pollutants from low sources and can lead to higher pollutant concentrations at night and in the morning hours. Because of the frequent inversions in the region, the frequency of fog during the cold half-year period is also greater – average annual number from 30 to 50 days, while the foggiest months are December (9-10 days) and November (5-6 days). The second microclimatic region “Sofia City center” covers the city’s most densely built-up part and is of 530-580 m altitude. Here, the city’s warming influence is greatest and the socalled “city heat island” is present. That is why air temperature here is considerable higher than in the “Sofia field”, as the difference for the absolute minimum temperature is around 6-7°C and for the average temperature for the coldest month – January – it is around 1°C. Here, the influence of the city on microclimate is well expressed, i.e. the influence of heating of the city is greater when the minimal air temperature (Tmin) in the open field is lower. Thus, for example, 42
Ecology of the Sofia City CyanMagentaYellowBlack Page 43
Climate and Human Comfort of Sofia
when in the field Tmin= -10°C, in the city it is 1-2°C warmer, and at Tmin= -30°C, in the city it is 6-8°C warmer on average. The frequent temperature inversions in the Sofia hollow valley lead to an increased number of fog days within the city in winter, reaching an average of 35-40 days for all of the three winter months, with maximum in December (13-15 days). The cause of this is the intensified pollution of city air and that is why fog is most frequent at the lower locations of the central city parts, i.e. the Hadji Dimitar district, around Vazrazdane sq., Stochna gara, the Central railway station, etc. It is characteristic that fog is longer-lasting in the city than in the open field. This creates conditions of higher levels of air pollution. Fog in urban and industrial areas, where considerable amounts of SO2 are emitted, create “sulphur-acidic (winter) smog”, and the fog droplets contain toxic compounds. The centralization of home heating decreased air pollution in the city of Sofia and especially in its central parts. After its large-scale introduction, the frequency and the duration of fog diminished considerably and, respectively, the level of air pollution dropped. The third microclimatic region “Peripheral districts of Sofia” or “Peripheral city area” covers the districts around the center. The region can be separated into two parts: low (520-550m altitude) and high (550-580m). The low part encompasses wholly or partially the districts of Gara Iskar, Hristo Botev, Vassil Levski, Suhata reka, Malashevci, Orlandovci, Voenna rampa, Ilijanci, Nadezda, Momkova mahala and Vrabnica. By location, as well as by microclimate, this zone takes a transitional position between the hollow field and the city centre. The low terrain location determines the great frequency of temperature inversions and, therefore frequent fogs (30-40 days on average in winter months). The percentage of calm weather or winds up to 1m/s is high (about 65% of the days). This weak ventilation, frequent fogs, and the presence of many industrial plants cause high levels of air pollution. This can be seen from the average concentrations of pollutants, and the level is very high during calm weather. The high part of the peripheral city region covers wholly or partially the districts of Zapaden park, Razsadnika, Ivan Vazov, Hipodrouma, Iztok, Geo Milev and Slatina. Here conditions are more favourable for living compared to the low part. This relates both to micrometeorological conditions (greater altitude, better planting, proximity to the Vitosha mountains for some districts) and to air purity (lack of big industrial polluters). Here the ventilation is improved (windless days are 3040%), fog is rarely formed (20-30 days in winter) and has a shorter duration and lower intensity. At these conditions, the level of air pollution is lower in the high part compared to the low part of the city peripheral region. The fourth microclimatic region “Foot of the mountains” can also be divided into two areas. The low part (580-600 m altitude) includes the districts of Pavlovo, Buckstone, Ovtcha kupell, Borovo, Juzen park, Losenec, Ljulin (the southern parts), Mladost (the northern parts), and Gorubljane. The middle-high part (600-700 m altitude) includes districts and near-city villages like Bankja, Ivanjane, Suhodol, Gorna Banja, Knjazevo, Bojana, Dragalevci, Simeonovo, Hladilnika, Dianabad, Student city and district “Mladost” (the southern parts). These are parts of the Sofia valley with less densely built-up areas, and with green areas located at the foot of the Vitosha, Ljulin and Lozenska mountains, so the adverse influences of the city complex is significantly reduced. These parts are characterized by rapid horizontal changes in the values of meteorological elements that allows the division of the region into the above-mentioned 43
Ecology of the Sofia City CyanMagentaYellowBlack Page 44
V. Andreev et al.
two parts. For example, the frequency of fog in the winter is 20-25 days in the low part, compare to only 7-8 days in the highest part. The intensity of fog here is also considerably lower compared to other microclimatic regions of the city of Sofia. The value of the total solar radiation during winter is higher, getting up to 15-16 kcal/cm2 as compared to values of 10-13 kcal/cm2 in the central parts of the city. This is a result of the higher air transparency in the considered region. The decreased number of days with fog is also an indicator of better air quality. Therefore, the living conditions in this region are more favorable compared with that in the other regions of the city from both a microclimatic (higher temperatures in winter, cooler days in summer, higher intensity of the solar radiation, decreased number of days with fog) and an air quality point of view (lack of big industrial enterprises, better ventilation). The presence of local mountain-valley wind during calm weather (anticyclonic or low pressure gradient) contributes to better air quality, too. This wind usually blows in the night and early in the morning as it is a light (velocity of 0.5-3.0 m/s), fresh and from the south (S, SSE, SE, SW). It is observed mainly in August and September. Conventional measurement instruments have their peculiarities, transforming into weaknesses in some specialized studies. Thus, for example, the balloon radiosounding is not appropriate for studying the diurnal evolution of the atmospheric boundary layer because of the relatively high inertia of sensors at relatively high elevation speeds of the balloon. We have also noted that the sensors’ inertia requires a comparatively slow elevation of an observation balloon, and an appropriate method of time interpolation is needed for calculating data towards the same moment in time. That is why for some studies the use of other measurement methods is appropriate. An example is the complex experiment in Sofia in the summer of 1994 using a lidar system for determining the concentration of aerosol particles in air together with equipment for direct meteorological measurements (a universal ground station, an observation balloon, pilot balloons). The aim was, with frequent measurements of the spatial distribution of aerosol particles and profiles of temperature, humidity and wind, to study the diurnal evolution of the atmospheric boundary layer in general, and also over different ground surfaces in the city. The lidar and meteorological equipment was based in the southeastern part of Sofia – the Institute of Electronics and NIMH-BAN. Here, the route for lidar sounding started with a 900 m2 of grass cover area and then it moved to a residential area with 8-10 stories high blocks of flats (Mladost 1A). The forming of a stably stratified boundary layer at night, which was destroyed after sunrise, when the formation of a convective boundary layer (also called mixing layer) begins, represented considerable interest. The two-dimensional lidar schemes of the aerosol stratification in Fig. 12 show that the altitude of the mixing layer over the park zones is lower than over the residential areas (respectively around 100 and 180 m around 1 hour after sunrise and up to around 260 and 470 m at noon). Some useful results were also obtained with the adapted thermodynamical model of Stull (see Donev et al., 1995). For example, for the vertical heat flux that causes the growth of the mixing layer, a value of 0,15degK.m/s is obtained over the earth surface in residential areas. The average speed of growing of the mixing layer was determined for the period of measurement (from 07:30 to 11:10 hrs) – around 45 m/h over the park zone and 85 m/h over the residential area. The probable cause for this is the difference in intensity of the heat flux. 44
Ecology of the Sofia City CyanMagentaYellowBlack Page 45
Climate and Human Comfort of Sofia 30 June 1994, 07:20-07:30 LST, Az 220°
30 June 1994, 10:10-10:20 LST, Az 220°
300 200 100 0 0
200
400
600 800 1000 RANGE (m)
1200
1400
10 9 8 7 6 5 4 3 2 1 0
500 HEIGHT (m)
HEIGHT (m)
400
400 300 200 100 0
0
30 June 1994, 08:05-08:15 LST, Az 220°
200 100 0 400
600 800 RANGE (m)
1000
1200
1400
10 9 8 7 6 5 4 3 2 1 0
600 800 1000 RANGE (m)
1200
1400
500 HEIGHT (m)
HEIGHT (m)
300
200
400
30 June 1994, 11:00-11:10 LST, Az 220°
400
0
200
10 9 8 7 6 5 4 3 2 1 0
400 300 200 100 0
0
200
400
600
800
1000
1200
1400
10 9 8 7 6 5 4 3 2 1 0
RANGE (m)
Fig. 12. Schemes of the vertical distribution of aerosol particles (concentrations are given in grayscale) by which the altitude of the mixing layer over the park zone (broken line) and over the living areas (broken line with dots) is determined.
The results concerning the interaction of the mountain-valley circulation and the city heat island allow (see Kolev et al., 2000) for the building of a physical model (Fig. 13) for the local air flows and for the structure of the boundary layer over a city, situated in a complex orography region such as the Sofia valley. It is flanked on the south by the Vitosha mountains (altitude around 2200m and an average slope of 100) and on the north by the Stara planina mountain (altitude around 1500m and an average slope of 50). At the developed night scheme of the mountain-valley circulation, the cold air flows moving down the mountain slopes begin to fill the volume of the valley and form a “lake” of cold air and its depth increases during the night. The warmer air of the “residual layer” remaining from the daily (convective) mixing layer is left over the lake of cold air. The descent of the cold layer down 3 Potential Temperature Profile
2
Capping Inversion Residual layer
Heat island Cold Pool
Mountain winds
Urban plume 1
Fig. 13. A scheme of the model of local air flows and the structure of the boundary layer at night over Sofia City at air-mass weather; 1-Lidar system (three-rayed aerosol lidar or Raman lidar); 2-observation balloon; 3pilot-balloon (see Kolev et al., 2000).
45
Ecology of the Sofia City CyanMagentaYellowBlack Page 46
V. Andreev et al.
the slopes continues until the lake reaches an altitude where the air has the same temperature. In this way, the so called “valley inversion” is most frequently formed, which is a stable stratification in the volume of the lake (Whiteman, 1982). The mountain wind gradually destroys the city heat island and a ground inversion is formed. The result is that air flow from opposite directions is formed at some altitude over the city. It transports warm air and aerosol particles from the central part of the city heat island and creates the so-called “urban plume”. In the air above the city, between the urban plume and the residual layer, the first layer of an upper inversion is formed (see Potential Temperature Profile in the right part of Fig. 13). The second upper inversion is called “capping inversion” because it capped the mixing layer during the day. The complex lidar and direct meteorological measurements allowed us to form general schemes for the typical vertical profile of the potential temperature and the characteristic altitudes of the layers with an increased concentration of aerosol particles over Sofia City during the experiment in 1994 (Fig. 14). These are typical for air-mass synoptic situations. The first characteristic layer is the ground temperature inversion with a thickness of around 100 m. It is formed due to radiation cooling of the earth surface and the surface air after sunset and the advection of cooler ambient air in the valley at night. This advection causes compensational fluxes in height, leading to the formation of a layer of warmer air over the ground inversion, called the urban plume of warm air. The plume transfers air from the central city parts (with a greater concentration of aerosol particles in it) and the potential temperature is almost constant with height. Thus, between altitudes of 200 and 400 m, the first aerosol layer is formed due to the retention of aerosol particles by the first ground inversion, situated in the layer between the urban plume and the residual layer, most frequently between 300 and 400 m. The aerosol particles that are emitted in the air from sources at the earth surface, elevate in height by means of the diffusion 1000
Height above surface (m)
900 800
Aerosol layer
Capping inversion
700 600
Residual layer
500 400
Aerosol layer
Lifted inversion
300 200
Urban plume
100
Ground based inversion
Cold Pool
0 6
8
10
12
14
16
18
20
Potential Temperature (°C)
Fig. 14. Typical profiles of the potential temperature and altitude of the aerosol layer disposition over Sofia City, obtained by complexion of lidar and direct meteorological data during short-term experiments.
46
Ecology of the Sofia City CyanMagentaYellowBlack Page 47
Climate and Human Comfort of Sofia
processes in the mixing layer forming during the day. After sunset, a portion of these elevated particles stay at the residual layer, under the capping inversion, its base being situated at around 700 m. It must be noted that the values of the parameters of the inversion and aerosol layers depend on the meteorological situation and the season of the year. As was noted, the local wind field can be deformed over complex relief areas. We do not have an idea of the local wind field and its deformations in different synoptically defined winds. This is because, in practice, it is not possible to carry out a meteorological network with regular observations by automatic or conventional instruments that is dense enough for registration of microscale processes. Such direct observations in the natural atmosphere are rarely organized in the frames of special projects by temporary networks. These can, however, be replaced with surrogates such as laboratory (analogue) and numerical (computer) models. We were introduced to the results of Tucker (1989) that refer to laboratory modeling of the local flows over the Sofia valley. He used the water tank in which a model of the relief of the valley is towed to represent air flow from each of the four cardinal directions. Visualization of the water flow was achieved by emitting dye from eight small tubes placed around the valley and one in the industrial town of Pernik. The dye was emitted at a height which corresponds to 400m above the valley floor. Neutral buoyancy for the dye is obtained by drawing fluid from the level of the exit tubes. Vertical stratification of the water is obtained by adjusting the proportion of fresh to saline water as the tank is being filled. This study refers to the atmospheric situations characterized by stable conditions and light but well-defined prevailing winds in the valley. The laboratory experiments show the following. When northerly winds occur, flows from the western and central parts of the Sofia valley (e.g. from directions of Dragoman and from Sofia City respectively) converge onto the Pernik area where a lee eddy sometimes occurs. There are short periods of blocking of the circulation in the area to the south-west of Sofia before the fluid moves on southwards. In periods with easterly winds, blocking occurs over Sofia, while an eddy in the lee of Vitosha mountain causes a local circulation of air over Pernik; occasional insertions of air from Sofia can be observed. In the case of southerly winds, flow over the western half of the Sofia valley, the fluid moves northwards over the low hills and north-westwards out of the valley. The central and eastern parts of the valley are prone to blocking. Lee eddies occur at various times and the air appears to be trapped over the whole of the eastern half of the valley. In situations with westerly wind, the main flow is formed by fluid entering the valley from the northwest, near the town of Dragoman, and moves south-eastwards across the northern part of the valley. It is interesting that a flow from the Pernik area moves north-eastwards into the Sofia valley and then moves south-eastwards, staying on the southern side of the valley. In the eastern part, these two streams converge and flow down the Luda Yana and Topolnitza valleys, probably causing a zone of stagnation in the upper part of the Maritza valley. The quasi-hydrostatic meso-scale model, developed by Ganev K. et al. (see Ganev, 1993) has been adapted for the Sofia valley (Dimitrova & Ganev, 2000) and used for determining the meteorological fields. The adaptation was made by means of numerical experiments with a twodimensional version of the model and their comparison with measurements made during laboratory experiments in an aerodynamic tube. 47
Ecology of the Sofia City CyanMagentaYellowBlack Page 48
V. Andreev et al.
The modeling calculations are conducted at the local scale (dimensions of region 105 and 75 km and 1.5 km step at 20 vertical levels) at the following stationary meteorological conditions: windless or different directions of a given weak background wind; stable and unstable stratification of the air temperature. The modeling results in conditions of stable stratification coincide qualitatively with conclusions for the local air flows from laboratory water tank experiments (Tucker, 1989) and with air pollutant distributions from field campaigns (Andreev, 1998). Fig. 15 shows the fields of the simulated wind at a height of 100 m in the Sofia valley at northern and southern background winds. It can be seen that considerable parts of the valley are occupied by flows in the opposite direction to the given background wind. With a lack of background synoptic wind, the slope and mountain-valley winds have a determining role in the creation of air circulation at local areas such as the Sofia valley (Fig. 16) at both stratifications. a) 5 m/s
b) 5 m/s
Kostinbrod
Kostinbrod Kremikovtzi
Kremikovtzi
Sofia Pernik
Sofia Pernik
Fig. 15. The fields of simulated winds in the Sofia valley at 100 m altitude and given (a) north and (b) south background synoptic winds and a stable temperature stratification. a) 5 m/s
b) 5 m/s
Kostinbrod
Kostinbrod Kremikovtzi
Kremikovtzi
Sofia Pernik
Sofia Pernik
Fig. 16. The fields of simulated winds in the Sofia valley at 100 m altitude with no background synoptic wind and (a) stable and (b) unstable temperature stratification.
48
Ecology of the Sofia City CyanMagentaYellowBlack Page 49
Climate and Human Comfort of Sofia
In some of the numerical experiments, the influence of factors such as wind speed, cloud cover and the anthropogenic heat flux from city sources was observed on the formation and the diurnal development of the city heat island (Dimitrova, 2000). The greatest contribution comes from the air humidity and the heat conductivity of the earth surface, and the contribution of roughness is negative. Wind and dense cloud cover have the greatest influence of all atmospheric parameters. To evaluate the human comfort conditions in Sofia and its surroundings, data from the meteorological stations Sofia-NIMH, Gorni Lozen, Simeonovo, Pancharevo, Buhovo and Busmantsy were used from 1990 onwards when a maximum number of stations worked simultaneously; this period is 1990-1994. After 1994, stations Simeonovo, Buhovo and Pancharevo were closed. Data are for the climatic terms 07, 14 and 21 hrs for the months July and January, which are representative for warm and cold half-year periods. In order to obtain assessment of climate comfortableness in the region of Sofia, the cooling index K was used, calculated using formula (1), through which, according to Siple & Passel (1945), different classes (types) of weather can be introduced (Table 1). The most favourable class of weather from human perspective is the neutral class when the human body neither looses, nor receives heat. The neighbouring classses of weather can still be considered comfortable and close to the comfortable types of weather. The hypotonic type corresponds to heat energy reception weather that can be regarded as warm subcomfort, and the tonic weather type is favourable for sporting activities. The remaining weather types are uncomfortable. Thus, for example, at the cooling weather type, the human body cools down due to loss of heat. The endothermic and atonic types are also uncomfortable because then the human body receives heat. Here it should be noted that for sea resorts the preferable weather type by tourists is atonic. In mountain resorts during the skiing season, the cooling weather type is favourable. In winter, naturally, endothermic and atonic types of weather are not observed (Table 2). The most favourable weather type is therefore the neutral comfortableness class (K=3). The frequency of this class is greatest in the region of Gorni Lozen, followed by Mladost 1 NIMH. The tonic class (K=4), i.e. cool subcomfort suitable for normal activity in the open, is most frequently observed in the area of the village Buhovo (62%) and Busmantsy (53%). The cold discomfort or the cooling weather class prevails also in the area of Busmantsy (47%) and Buhovo (30.8%). Table 1. Definition of the bioclimatic weather types (classes) by means of the cooling index K Type (class) weather Endothermic Atonic Hypotonic Neutral Tonic Cooling
(K=0) (K=1) (K=2) (K=3) (K=4) (K=5)
Values of the cooling index K [W/m2] 0< 175 ≤ 350 ≤ 700 ≤
K≤0 K ≤ 174 K ≤ 349 K ≤ 699 K ≤ 1049 K > 1050
49
Ecology of the Sofia City CyanMagentaYellowBlack Page 50
V. Andreev et al. Table 2. Frequency percentage of the different bioclimatic weather types in the Sofia valley (% in winter / % in summer). The symbol “∑” means the sum of the percent of the types with K equal to 2, 3 and 4. Type (class) station NIMH Gorni Lozen Simeonovo Pancharevo Buhovo Busmantsy
K=0
K=1
K=2
K=3
K=4
K=5
Σ
-
- /39.1 - /21.3 - /18.1 - /26.5 - /8.6 -/18
12.1/40 9/44.5 - /54.3 2.4/29.4 - /29.8 - /77
47.3/20.4 50.2/33.3 30.6/37.6 36.6/44.1 7.2/60 - /5
26.1/ 28.7/ 46.8/ 31.7/ 62.0/ 53/ -
14.5/ 12.2/ 21/ 19.3/ 30.8/ 47/ -
85.5/60.4 87.9/77.8 77.4/91.9 70.7/73.9 69.2/89.8 53/82
Interestingly, at some stations separate occasions of warm subcomfort are observed (hypotonic class, K=2), probably connected with periods of warming weather. This class of weather is quite favourable, especially in the cold half-year period. To determine the degree of human comfort, it is better to summarize the frequency percentage of classes with K 2, 3 and 4 as shown in Table 2 with the symbol “∑”. With this assessment, the living conditions in winter are again best in Gorni Lozen and Mladost 1-NIMH. Living conditions in winter, using this indicator, are close to good in the regions in close proximity to Simeonovo and Pancharevo. The conditions are most uncomfortable in Busmantsy and Buhovo where the cold discomfort prevails due to the openness of the terrain and the conditions for forming of low temperatures. It must be noted that by using this summary indicator, the built-up part of Druzba as a whole has relatively high comfortableness conditions due to higher air temperatures and weaker wind. In summer (Table 2), naturally, the cold discomfort – the cooling comfortableness class (K=4) – is absent. Comfortable conditions are present most frequently in Buhovo – 60%, and most rarely in Busmantsy (5%). The percentage of this class is relatively high at Pancharevo (44%), Simeonovo (38%) and Gorni Lozen (33%). The hot discomfort (K=1) prevails in the city parts – NIMH station – 39%, and is most rarely observed in Buhovo, Busmantsy and Simeonovo. The hot subcomfort (K=2) is present everywhere, when in spite of the relative overheating of the organism, the conditions are still favourable for activities in the open. For this season, an assessment is also made for the human comfort conditions, based on the summary frequency of classes K=3 and K=2. According to this assessment, the climatic comfort conditions are most favourable in Simeonovo and Buhovo. So, the summary indicator at Busmantsy is relatively high (82%) due to the high percentage of hot subcomfort (77% of hypotonic class, K=2) in spite of the low frequency of the comfortable class (neutral type, K=3). The sum of classes with K equal to 2 and 3 is favourable for Gorni Lozen (78%) and Pancharevo (74%).
CONCLUSIONS Local conditions and especially the relatively high altitude, relief and city complex, enhance continental features of Sofia City’s climate. One of the most characteristic features of the city climate – “the heat island”, is found here too, but its intensity is studied only as the 50
Ecology of the Sofia City CyanMagentaYellowBlack Page 51
Climate and Human Comfort of Sofia
difference (∆Τ) between temperatures in singular urban, suburban and rural stations. The average value of ∆Τ is ≥ 3°C with highest values reaching up to 10°C. An idea of the form of the heat island is given by experimental studies, but the small number of stations within the city does not allow us to study its structure thoroughly, including determining the presence, number and parameters of the secondary islands. A secondary heat island is detected in the Druzba district (see Fig. 11). The influence of wind speed and cloud cover on the difference in ∆Τ is also studied. Wind change in different parts of the city is obtained by data from short-term experiments and mobile measurements. The spatial distribution of the wind change coefficient Kv confirms the reducing influence of the city on wind speed. The presence of slope wind from the Vitosha mountains was detected in the southern districts of the city at high pressure fields, including small gradient ones. Wind appears mainly at night during the 17-07 hrs period. This is favourable regarding air pollution, because clean mountain air enters the city by such winds. An important result is that high buildings can stop this wind in the periphery of the city. This is an important weakness of the city planning of Sofia. A very useful result from climatic studies in cities is the implementation of a reliable microclimatic regioning. It is done on the basis of relatively long series of observations (e.g., for least a 10year period) at a sufficiently large number of stations in diffrent regions of the city. Fortunately, such series of observations are available for a few climatic stations in Sofia, though most were closed during the last 10-15 years. Regioning divided Sofia into four microclimatic regions. For each region, the characteristic values and changes throughout the year of the basic meteorological elements as well as air quality and human comfort values are given. This microclimatic regioning should be the basis for city planing. Unfortunately, this is not the case for any town in Bulgaria, including the capital. When determining the location, the prevailing orientation of buildings and some other characteristics of the big complexes in Sofia, other considerations than the determining role of climate and microclimatic peculiarities of the city have been taken into account. Only in this way can we explain why most of the city complexes are situated not only in the industrial, but also in the lowest parts of the city, from where it naturally follows that they are the regions with highly polluted air and with a relatively low degree of comfortableness. At the same time, large parts of the Sofia valley and Sofia City situated at the foot of the Vitosha mountains, Ljulin, Lozenska, and even Stara planina, remain uncultivated. From this point of view, the big complexes Mladost and Ljulin have a more suitable location. There, though, other planning defects with respect to climate and microclimate exist. Thus, for example, it is not known why the main part of the apartment blocks in Mladost have a west-east orientation, i.e. against the prevailing winds. A decisive factor for this was probably the use of the already existing main street, now called Al. Malinov Bldv. It was possible, however, for the orientation to be south-north as it would lead to better inner comfort of the blocks. Then the heat regime of the buildings would have been improved considerably. Now, however, heating is low due to the strong cooling influence of the frequent west, north-west and east winds. In the complexes close to the mountains, i.e. the mountain Vitosha, other planning errors were allowed: very high and very wide apartment blocks were built at the periphery of the complexes towards the mountains. Besides, there are no boulevards and streets that start from the mountains 51
Ecology of the Sofia City CyanMagentaYellowBlack Page 52
V. Andreev et al.
and enter radially towards the center of the town. These routes would ensure greater penetration of mountain winds that bring clean mountain air into the city at night, and also coolness in summer. Overcoming these city planning defects is one of the main goals of the prognosis part (until 2020) of the “General Structural Plan of Sofia” (2002). The greatest advantage to be achieved is to divide the city into regions based on a complex characterization of the separate city parts such as microclimate, climatic comfort conditions and levels of air pollution with harmful compounds. Some of the authors of this publication completed a complex regioning on two indicators – microclimate and level of air pollution at the request of a construction company (still unpublished) – see the article of Andreev et al.: “Air Quality of Sofia – Meteorological Aspects” (this book).
ACKNOWLEDGMENTS We thank the editors L. Penev, J. Niemelä, D. J. Kotze and N. Chipev for the invitation to participate in compiling this interesting collection and for final editing of the manuscript.
REFERENCES Andreev V. 1980. Main Results from the Complex Experiment Sofia’76. Hydrology and Meteorology, 1: 7-15. Andreev V. 1998. An Attempt to Explain Some Peculiarities of Sofia Air Pollution by the Laboratory Flow Modelling. Bulg. J. of Meteorol. and Hydrology 4: 200-209. Andreev V. 2001. Aviation Meteorology. Textbook for students of Technical University. Sofia. 346 pp. (in Bulgarian). Batchvarova E., S.-E. Gryning. 1991. Applied Model for the Growth of the Daytime Mixed Layer. BoundaryLayer Meteorol. 56: 261-274. Batchvarova E., D. Syrakov, A.Tzenkova. 1996. Air Pollution Characteristics of a Region of Sofia and Data from Field Experiments (1992-1993). In: Urban Air Pollution, Monitoring and control strategies(I. Allegrini and F. De Santis. Eds.) pp. 235-242. NATO ASI Series, 2. Environment – v. 8. Besancenot, J.P. 1978. Le bioclimat humain de Rio. In: Recherches de Climatologie en Milieu Tropical et Mediterranean. (Suchel J.B., E. Altes, J.P. Besancenot & P. Maheras Eds.). Cahier No. 6 du Centre de Recherches de Climatologie, Universite de Dijon, Dijon. Blaskova D., G. Kurchatova, Tz. Tchapanov, V. Andreev, P. Kazasov. 1968. Regime of the Temperature Inversions in Sofia District and Their Influence on the Pollution of the Atmospheric Air by Sulpher Dioxide. Bulletin of the Institute of Hydrology and Meteorology Sofia, 13: 103-146 (in Bulgarian). Blaskova D, L. Zlatkova, S. Lingova, Zh. Modeva, L. Subev, M. Teneva. 1983. Climate and Microclimate of Sofia. Publ. BAS, Sofia. 154 pp. (in Bulgarian). Blaskova D, S. Lingova. 1991. Urban influence on climate. In: The Climate of Bulgaria. Publ. BAS, Sofia. 438450, (in Bulgarian). Branzov Ch., S. Panchev. 1991. One-dimensional Nonstationary Semiempirical Model of Radiative Temperature Inversion on Sofia. Bulg. J. of Meteorol. and Hydrology. 11(2): 9-17.
52
Ecology of the Sofia City CyanMagentaYellowBlack Page 53
Climate and Human Comfort of Sofia Branzov Ch., E. Donev, S. Panchev. 1992. Experimental Study of the Structure and Dynamics of Temperature Inversions and Convective Boundary Layer on the area of Sofia. Sofia University Annual 82: 147-166. Davis, N.E.1968. An optimum summer weather index. Weather 23: 305. Donev E., K. Zeller, St. Panchev, I. Kolev. 1995. Boundary Layer Growth and Lapse Rate Changes Determined by Lidar and Surface Heat Flux in Sofia. Acta Meteorological Sinica 9: 101-111. Dimitrova R. 2000. Numerical Studies of the City Heat Island. Bulg. Geophysic J. 26(2). Dimitrova R., K. Ganev. 2000. Numerical Modeling of Air Flows and Pollution Transport in the Sofia Valley. III National Conference in Geophysics, 11-13 Oct. 2000. Sofia, pp. 77-78. Ganev K. 1993. Numerical Study of the Local Flow Systems in the “Kozloduy” NPP Region. Bulg. Geophys.J. 19(1): 9-23. General Structural Plan of Sofia. Stage III. Summarized report. Part: Ecology. “Sofproekt”. May 2002. Gregorczuk, M. Cena, K. 1967. Distribution of effective temperature over the surface of the earth. Int. J. Biometeor. 11: 145-149. Houghten, F.C.& Yaglou, C.P. 1923. Determining lines of equal comfort. Trans. Amer. Soc. Heat. Vent. Engrs. 29: 163-76. Kandror I.S., Demina D.M., Ratner E.M. 1974. Physiological principles of hygienic and climate area division of the territory of the USSR. Moscow. (in Russian) Kondo J., H. Yamazawa. 1986. Aerodynamic roughness over an inhomogeneous ground surface. BoundaryLayer Meteorol. 35: 387-407. Kolev I., P. Savov, B. Kaprielov, O. Parvanov, V. Simeonov. 2000. Lidar Observation of the Nocturnal Boundary Layer Formation over Sofia, Bulgaria. Atmospheric Environment 34: 3223-3235. Koleva Ek. 1988. Peculiarities of the Precipitation Distribution at the flat part of Bulgaria. Problems of Meteorology and Hydrology 2: 41-48, (in Bulgarian). Koleva Ek. 1991. A Precipitation Distribution. In: Climate of Bulgaria. BAS Publ., Sofia. 499 pp. (in Bulgarian) Koleva Ek., L. Krastev, E. Peneva, E. Stanev. 1996. Verification of high resolution climatic simulations for the area of Bulgaria. Part I: The state of the climate for the period 1961-1990. Bulgarian Journal of Meteorology & Hydrology 7(3/4): 73-84. Kyle,W.J. 1994. The human bioclimate of Hong Kong. Contemporary climatology. In: Proceedings of the meeting of the Commision on Climatology of the International Geographical Union, 15-20 Aug 1994, BRNO, Czech Republic, pp. 345-350. Landsberg H. E. 1970. Man-made climatic changes. Science 170: 1265-1274. Marinov V. 1971. Physiological hygienic and climate area division of the territory of Bulgaria. Sofia, (in Bulgarian). Mestayer P., S. Anquetin. 1995. Climatology of Cities. In: Diffusion and Transport of Pollutants in Atmospheric Mesoscale Flow Fields. A. Gyr and F.S. Rys (Eds). Kluwer Academic Publishers. pp. 165-189. Myrup L., C. McGinn, R. Flocchini. 1993. An analysis of microclimatic variation in a suburban environment. Atmospheric Environment 27B: 129-156. Oke T .R. 1990. Bibliography of urban climate 1981-1988. WCAP-15, WMO/TD No397, 62 p. Oke T., G. Johnson, D. Steyn, I. Watson. 1991. Simulation of surface urban heat islands under ‘ideal’ conditions in night. Part 2: Diagnosis of causation. Boundary-Layer Meteorol. 56: 339-358. Palmgren F. 1996. Design and Planning of Field Campaigns. In: Urban Air Pollution. Monitoring and Control Strategies.(I. Allegrini and F. De Santis. Eds.). pp. 77-99. NATO ASI Series / 2, Environment, v. 8. Rusanov V.I. 1973. Methods of investigation of climate for medical purposes. Tomsk. (in Russian)
53
Ecology of the Sofia City CyanMagentaYellowBlack Page 54
V. Andreev et al. Siple P.A., C.F. Passel. 1945. Measurements of dry atmospheric cooling in subfreezing temperatures. Proc. Amer. Phil. Soc. 89: 177-199. Syrakov D (Leader of a Scientific Team on Contract No H3-46/91 with National Fund “Scientific Investigations”). Theme “Methods for a Complex Meteorological Expertise of the Air Pollution”. Reports on I stage-1992, II stage-1993 and III stage -1994 (in Bulgarian). Syrakova M., M. Zaharieva. 1998. Sofia Heat Island-Diurnal and Seasonal Variations. Part I: Frequency Distribution of Hourly Temperature Differences. Bulg. J. of Meteorol. and Hydrology. 14: 210-219. Syrakova M., M. Zaharieva. 1999. Sofia Heat Island-Diurnal and Seasonal Variations. Part II: Diurnal and Seasonal Variations of the Heat Island Intensity. Bulg. J. of Meteorol. and Hydrology, 11: 1-16. Temnikova N. S. 1975. Theoretical basis of recreation geography. Moscow. (in Russian). Teneva M., E. Batchvarova. 1989. Experimental Investigation of Urban Air Pollution. Bulgarian Geophysical Journal, 15(13): 62-69 (in Bulgarian). Teneva M., E. Batchvarova, L. Iordanova. 1989. Air Pollution Peculiarities in Complex Orography Areas. In: Proceeding XIV International Conference on Carpathian Meteorology, pp. 396-401. 25-30 Sept 1989, Sofia, Bulgaria (in Russian). Terjung, W.H. 1966. Physiologic climates of the coterminous United States: a bioclimatic classification based on man. Ann. Ass. Amer. Georgs. 56: 141-179. Terjung, W.H. 1968. World patterns of the distribution of the monthly comfort index. Int. J. Biometeor. 12: 119-151. Terjung, W.H. & S.F. Louie. 1971. Potential solar radiation climate of man. Ann. Ass. Amer. Geogrs. 61: 481-500. Thom, E.C. 1959. The discomfort index. Weatherwise, 12: 57-60. Tucker G.B. 1989. Laboratory Modelling of Air Flow in the Sofia Valley. Bulg. Geophys. J. 15(1): 92-100. Tzenkova A., J. Ivancheva. 1999. Microclimate Characteristics of the Urban Area. 4th European Conference on Applications of Meteorology, ECAM’99, 13-17 September 1999, Norrkoping, Sweden. TzenkovaA., J. Ivancheva. 2000. Human comfort in Sofia and its surrounding area. In: 3rd European Conference on Applied Climatology, 16-20 Oct., Pisa, Italy, CD. Whiteman C. D. 1982. Breakup of Temperature Inversion in Deep Mountain Valleys: Part I. Observations. Journal of Applied Meteorology 21: 270-289. Zlatkova L., A. Tzenkova. 1985. Urban Microclimate Influenced by Air Pollution. In: Modelling, Analysis and Standartisation of Air Pollution (Scientific Report, Institute of Hydrology and Meteorology), Sofia (in Bulgarian).
54
Ecology of the Sofia City CyanMagentaYellowBlack Page 55
Urban Climate and Air Quality - a Review
Urban Climate and Air Quality - a Review Vassil M. Andreev National Institute of Meteorology & Hydrology; 1784 Sofia Fax: (+3592) 988 44 94; E-mail: [email protected]
SUMMARY Humans in urban areas are exposed to the simultaneous influence by agents of meteorological, geophysical and air quality origins. Urban climate and air quality are interesting problems as people all over the world are highly urbanized. A review of the knowledge of these agents is offered in this article. The following problems are considered: in-canopy energy budget, urban heat island and local convective circulation; the structure of the “urban boundary layer” in conditions of different wind speeds; the influence of the city on the atmosphere over the downwind regions and on cloud and convective precipitation processes, synoptic wind speeds and other weather conditions over the urban area. The major factors (sources of harmful emissions; meteorological conditions; chemical transformations) which determine air pollution in any region are also considered. The paper also focuses on: episodes with high air pollution, including winter and summer smog; Air Quality Guidelines and measures for the preservation of favourable air quality in urban areas and the necessity for an air quality monitoring network; meteorological data; emission inventories and numerical models for urban air pollution management are also discussed. Detailed predictions and real time control of air quality demand the development of complex sub-mesoscale models to understand the relations between the physical-geographical conditions, regional meteorology, local pollutant sources and the monitoring measurements.
KEY WORDS Urban climate, urban canopy, urban heat island, in-canopy energy budget, local convective circulation, urban boundary layer, air quality, air pollutants, air pollution sources, air pollution episodes, winter and summer smog, “hot spot”, Air Quality Guidelines, monitoring network, emission inventories, urban flow models, urban dispersion models, Sofia, Bulgaria.
L. Penev, J. Niemelä, D. J. Kotze & N. Chipev (Eds.) 2004. Ecology of the City of Sofia. Species and Communities in an Urban Environment, pp. 55-82 © PENSOFT Publishers Sofia-Moscow
55
Ecology of the Sofia City CyanMagentaYellowBlack Page 56
Vassil Andreev
INTRODUCTION Humans are exposed at every moment to the simultaneous influence of both meteorological (temperature, air humidity, wind speed, etc.) and geophysical (solar radiation, Earth magnetism) impacts. In urban and industrial areas, air quality is a third important agent. The mutual action of this agents is not equivalent to the sum of their separate actions. Knowledge of these agents is important because of an urbanized world. For instance, more than 70% of Europeans are living in urban areas today. The structure of the lower urban atmosphere and the local air quality in large cities are resulting from the interactions of processes taking place at three different scales. Most meteorological processes can be described at the meso-scales (atmospheric boundary layer height – several tens of km) although meteorological forcing is known to be due to larger scales. Urban ground is covered with natural and artificial materials, buildings and other obstacles. The heat budget as well as the flow structure close to the ground depend largely on the height and density of these constructions. Flow structure varies from a series of turbulent wakes due to isolated obstacles, to strongly interacting turbulent streaks and even to turbulent flows having some similarities with flows in semi-porous media. The atmospheric layer below roof level is called the urban sub-layer or urban canopy (see Fig. 1a), by analogy with the atmospheric layer between ground and trees of dense forest. Kondo and Yamazawa (1986) defines its height by the integral of the building heights: h= (∑Hi.Si)/S, where S is the total area under consideration, Hi and Si are the height and area of the individual obstacles (buildings). Many dispersion processes are studied in the immediate vicinity of the sources, or at the impact targets, and therefore must be studied at small scales within the urban canopy; this is especially true for those emissions of pollutants which transform rapidly and interact chemically. Most of the pollutants of interest for evaluating the impacts on human health, vegetation, materials, monuments or on the long term atmospheric content (global change) are secondary compounds, not directly produced by anthropogenic sources. So, when it is considered that urban dispersion problems must be regarded most of the time as the dispersion of photo-chemically transformable pollutants, it is realized that an intermediate range of processes must be studied. Their scales are those of the turbulent mixing of the not-so-rapid chemical transformations, and of the interactions between the source layers and the atmospheric boundary layer, because the dispersion processes of photochemical compounds depend on the dynamic, physical, micro-physical, chemical and radiative structures of the neighboring atmospheric layers. These are at the scales of the local convection movements that are specific to the urban environment and are due to local, even weak orographies combined with strong heterogeneities in the surface temperatures and in the canopy dynamic structure (Fig. 2). These scales are called sub-meso scale because they are smaller than the scales usually simulated by meso-scale models. Indeed, the meso-scale models are usually unable to simulate the fine decametric scales and are so heterogeneous in urban areas that it is not well known how to model them. Up to now, meso-scale models cannot correctly predict the pertinent urban patterns, even in combination with sophisticated photochemical models, because of their inability to simulate the necessary details of the air movements (Moussiopoulos et al., 1992). 56
Ecology of the Sofia City CyanMagentaYellowBlack Page 57
Urban Climate and Air Quality - a Review
URBAN CLIMATE The actual climate of urban areas are formed in the background of the regional climate typical of the geographical area but is deformed under the influence of the local relief and human-made fabrics. On the other hand, there is a complex connection between the elements of urbanization (a building, a street, a plant, a park, a parking lot, etc.) and the environment in which it exists. As a result, these elements or a complex of elements form their own, specific microclimate. That is why only a few micrometeorological rules are general enough to be exported from one city to another. The most studied climatologic effect of cities is the urban heat island, i.e. the systematic higher temperature of the atmosphere over the city compared to that of the upwind neighbouring rural area. Because there is no recent reference handbook on the urban meteorology in Bulgaria, following the paper of Mestayer & Anquetin (1995), we first survey the observations assessing the heat island, its traditional conceptual description, its causes and extents and the induced climatic perturbations at the regional scale, i.e. the city effects on the regional physics and microphysics of the atmosphere due to the altered fluxes of sensible heat and moisture, and the production of pollutant gases and particles: among them are the increases in cloud coverage, precipitation and fog.. Then we studied the consequences of heat island on the dispersion at meso and sub-meso scales, and on its simulation. We found that large problems remained in connection with the structure of the flows close to city roofs: this leads to a focus on the urban canopy, and a review of our knowledge of the in-canopy energy budget and pollutant dispersion. Finally we examine the present trends in climatology and dispersion at sub-mesoscales in the urban atmosphere. Heat islands have been detected in the XIXth century by European scientists examining historical records of large cities such as Paris, Berlin, Vienna and London (Landsberg, 1970). It appeared that the temperature differences between typical urban and rural stations ∆Tu-r increased with time and are dependent on the city size, from 20C for small towns up to 120C for very large ones. Measurements over large urban areas show that the internal patterns of the heat islands are rather complicated in space, as a result of the 3-dimensional influence of terrain heterogeneity. Usually, in big towns, there are many heat islands - a “main” one and a few – “secondaries”. They are also changing rapidly during the diurnal cycle and are dependent on the meteorological conditions with, for example, maximum levels appearing at the end of the afternoon or the beginning of the night in conditions of weak winds and clear air, and in winter (see Mestayer & Anquetin, 1995). Oke (1990) reviewed about 700 papers on this subject. The following physical processes have been considered as taking part in the heat island generation: anthropogenic sensible heat production due to house heating, vehicles and industries; reduction of wind speed due to the presence of buildings; alteration of the water balance due to the reduction of vegetation evapo-transpiration and to the water vapour directly released by industrial activities; alteration of the radiative budget, including the effects of greenhouse gas production; heat storage the constructions; local topography and local wind flows; alteration of the diurnal cycle of sensible heat flux, radiative heat flux, and conductive heat flux, especially the reduction of nocturnal cooling. The role and importance of these various processes appear to be dependent on the solar cycle and on season; e.g., the alteration of the water budget is a dominant factor in summer, while the dominant factor in winter seems to be the anthropogenic heat flux. A large number of measurements have been 57
Ecology of the Sofia City CyanMagentaYellowBlack Page 58
Vassil Andreev
taken to assess these effects in all types of cities, but most of the our knowledges are from three large experiments (see Mestayer & Anquetin, 1995):Vancouver, Canada (1973), St. Louis, USA (project METROMEX - Metropolitan Meteorological Experiment, 1974-1976) and Toulouse, France (1978). The first conclusion from these measurements is that the temperature differences can reach a very high value in a large and dense megalopolis. It is also interesting to note that even a small town of 1000 inhabitants, generates a noteworthy heat island. The structure of the town is also a factor of high importance, probably through the effect of ventilation that depends not only on the building density but also on building height. These studies have demonstrated that the heat island effect is important not only close to the ground but also higher in the atmospheric boundary layer and that it generates either “heat bubbles” or “heat plumes” (Fig. 1). During weak winds (or no wind), local circulations are formed because of the influence of the thermal convection caused from the heat island (Fig. 1b). This circulation consists of closed convective cells with an ascending branch over the heat island (the city center) and a descending branch over the suburbs. The surface branch of the convective cell brings clean and fresh air towards the warmer and polluted city parts. The aerodynamic model in Fig. 1a shows that in conditions of moderate to strong synoptic wind, an internal boundary layer inside a 2-D urban boundary layer over a flat relief develops, due either to a change in roughness (dynamic internal layer) or to a change in the surface heat flux (thermal internal layer). As a rule, these model flows are studied in wind tunnels.
Figure 1. Scheme of the urban heat island and the local air motions over an urban area situated at a flat relief in conditions of (a) weak wind and (b) moderate to strong wind. In the first case, the local convective type circulations are formed under the influence of the urban heat island (by Mestayer & Anquetin, 1995)
58
Ecology of the Sofia City CyanMagentaYellowBlack Page 59
Urban Climate and Air Quality - a Review
The schemes in Fig. 1a and Fig. 2a show the development of the urban boundary layer as one (or a few) internal layer(s). It shows that the city constructions are, essentially, a carpet of roughness elements and that the city landscape is a very complex terrain. Due to the induced city thermal and dynamic convection, the frequency and intensity of the inversion layers tend to decrease over the urban areas, and the “urban boundary layer” therefore tends to be, more often, close to neutral. The model of the structure of this boundary layer in Fig. 1 is a classical but crude approximation, corresponding to the pioneering modeling efforts of the 1970’s. Actually, all meteorological conditions produce some combination of the two situations in Fig. 1 and the picture is in fact more complicated: the structure of the “urban boundary layer” is 3-D, depends largely on the local orography (Fig. 2) and on the heterogeneity of the different districts within the city. Most of the studies of the last 25 years, however, have been developed within the framework of these conceptual models. Results from the three large experiments indicate the decreasing influence of the city on the synoptic wind speed. In summer, in the conditions of anticyclonic weather, the appearance of “slope wind”, also called “mountain (valley) breeze” is characteristic for the districts of urban areas, situated at the foot of a mountain (e.g. at the northern slope of Vitosha – the mountain near the city of Sofia). During the daytime and close to the earth surface, wind should blow from the city upwards on the mountain slope (anabatic wind), but it is very slight. The night slope wind, which is a katabatic type wind (as at Fig. 2) is well expressed, especially in the
Figure 2. Refinements of the scheme of the local air motions in Fig. 1 including an urban area situated in a complex relief. In this case (0), the local circulations are formed not only under the influence of the urban heat island but also as the effect of the local orography (the night “slope” wind) (by Mestayer & Anquetin, 1995).
59
Ecology of the Sofia City CyanMagentaYellowBlack Page 60
Vassil Andreev
middle of the summer in conditions of weak winds and cloudless weather. It begins to blow at the end of the afternoon (or the beginning of the night) and calms down shortly after sunrise. In clearly anticyclonic weather at night, this wind brings clean air, rich in negative ions from the mountain slope towards the polluted city parts. Because its vertical growth (up to 30-40m) and speed (1-3m/s) are minimal, the high buildings situated crosswise can stop the clean air at the periphery of the city. The observations show that the heat island appears to be a characteristic feature of the urban atmosphere and is evident over built-up areas of all sizes, even over small villages or isolated large districts, shopping centers, etc. The heat island parameters depend on the city size, structure, density, energy consumption and degree of pollution. Where pollution is low, the main factor seems to be the impact of solar radiation on the urban surfaces. Where pollution is high, energy consumption is also usually quite high and the anthropogenic energy flux seems to be the determining factor. It is, of course, always a combination of various factors. Oke (1973), on the base of experimental data, expressed the maximum urban-rural temperature difference as a function of the population size and of wind speed, as:
∆Tu-r = P 0,27/(4,04.Ur0,56)
(1)
where P is the number of inhabitants, Ur – the wind speed measured outside the city at a reference level of 10 m. It is obvious from this and other research (see Mestayer & Anquetin, 1995) that other factors need to be taken into account. For example, the structure of the heat island is considered to be of the “dome” type (Fig. 1b) for reference wind speeds lower than 3 m/s and of the “plume” type (Fig. 1a) for Ur>3 m/s. There are several parameterizations of the vertical extent of the heat island in Kimura et al. (see Mestayer & Anquetin, 1995): Hmax= a.R-L/6 where R= β.g.∆Tu.Lt 4/vt
(2)
where Lt is city length in meters; β – the coefficient of thermal expansion per unit volume; g – gravity; ∆Tu – the vertical temperature gradient over the urban area and vt the average turbulent diffusivity. Based on the St Louis METROMEX measurements, Auer and Changnon (see Mestayer & Anquetin, 1995) proposed a relationship between the heat island height and wind speed H= 1265 – 87,2 U for 2 ≤ U ≤ 8 m/s
(3)
The heat island is also determined by the development of the dynamic internal boundary layer when the wind is not weak. Its height is usually parameterized as a function of the fetch x, i.e. the distance in the change of roughness and of either the roughness lengths of the two surfaces z01 and z02 or that of the rougher surface z0M. The various formulae are shown in Mestayer & Anquetin (1995). They note that Jensen (1981) proposed to take into account the thermal stratification by parameterizing the internal layer height with Monin-Obukhov length scale L, as zIBL (x) = x3/2.L-1/2
for unstable stratification,
zIBL (x) = x1/2.L1/2
for stable stratification.
(4) All these parameterizations enter into the framework of the original 2-D scheme of the urban heat island. Some interactions of the heat island with atmosphere at the meso scale or with the 60
Ecology of the Sofia City CyanMagentaYellowBlack Page 61
Urban Climate and Air Quality - a Review
meso scale climate have been studied with a meso scale meteorological 3-D model. However, the meso-scale simulations considered relatively simple parameterizations of the canopy “terrain” that appear inadequate, particularly with respect to the parameterization of the surface sensible and latent heat fluxes, the subsurface heat storage and the role of urban geometry upon the surface energy budget. Most of our knowledge of the interaction of urban areas with the atmosphere comes mainly from natural experiments. The heat island in principle is related to the heat exchanges and energy budget within the canopy layer, and also to exchanges at the canopy layer - boundary layer interface, and therefore to the interaction between the meteorology and the canopy structure. At least six causes of heat build-up can be listed : the increase of long wave radiation from the sky; the decrease of solar radiation reflection due to radiative trapping; the increase of solar radiation absorption by urban materials due to their lower albedo; the increase of heat storage by the urban wall and ground materials; anthropogenic sources of heat and energy; and the decrease of latent heat consumption due to reduced vegetation and to the water collection networks. Four processes remain as possible causes of the heat island development in the urban boundary layer above roof level (see Fig. 1): the warm air provided by the canopy layer; the direct anthropogenic heat inputs by chimneys and high stacks; the entrainment of warm air by induced convective movements; and the increase of radiative transfers due to atmospheric pollution. It is interesting to analyze the difference in the atmospheric energy budgets over city surfaces and over rural areas, which can help us to understand the origin of a heat island. Following Mestayer & Anquetin (1995), the steady state energy budget over a flat surface can be written as follows: (Q+q)(1-a) + εItotal-εσT4=H+G+LE
(5)
where Q is direct solar radiation, q is diffuse solar radiation, a – surface albedo, ε – surface emissivity, Itotal – hemispherical long wave radiation, σ – Stefan-Boltzmann constant, T – surface temperature, H – sensible heat flux, G – conduction and LE – latent heat flux. This equation (5) can be rewritten to let the radiative budget appear on the left hand side and the heat budget on the right hand side of the equation. The budget over the urban canopy must include an additional anthropogenic source term F, essentially due to fuel consumption. In this case, it is usually signifying H and LE – the turbulent heat fluxes in the air, G – the ground-and-construction storage, R – the radiative budget, i.e. R+F=H+G+LE
(6)
If the urban canopy is considered a nearly horizontal surface, the radiative budget R can be simplified, i.e. R=(1-a).S+∆L=S-a.S+La-Ls
(7)
Here S is received (direct + diffuse) solar radiation, (a.S) – part of the solar radiation that is reflected by the canopy, La – long wave radiation emitted by the atmosphere downwards to the ground and Ls – long wave radiation emitted by the canopy upwards to the sky. Solar radiation S can be attenuated by 2 to 20% over urban areas. Air pollution is the main cause of this reduction. Short wave (solar) radiations are absorbed by small aerosols, water vapour, oxygen and ozone. A 61
Ecology of the Sofia City CyanMagentaYellowBlack Page 62
Vassil Andreev
part of the aerosols are condensation nuclei that increase the formation of clouds that also absorb short and long wave radiations. The part of the solar radiation that is re-emitted to the atmosphere by the urban canopy is essentially depending on the albedo of the urban surfaces. The albedo of a surface characterizes the visible radiation reflection of the surface; it is 1 for a surface that perfectly reflects all radiations, and 0 for a surface that totally absorbs all visible radiations. The albedo of a surface depends on its material, colour and surface state. It can vary with time, due to surface processes such as the deposition of dust or the elevation of the sun (glass surfaces have albedos varying from 0,08 to 0,5 as a function of solar incidence angle). Urban surfaces present a very large palette of albedos from 0,08 for tar and asphalt to 0,5 or 0,9 for white-painted walls. Most of these have albedos between 0,1 and 0,2, but, as a rule, urban albedos are lower than albedos over rural surfaces: the average urban albedo is about 0,13 while it is 0,16 over rural areas. As a whole, the canopy presents an even less effective albedo due to the radiative trap effect of street canyons and building yards: much incoming radiation undergoes multiple reflections on building walls and interspersed ground surfaces. Building height and density, and solar elevation influence both the radiative trap effect and the shadowing effect of one building on another. Oke (1988) showed that the effective albedo of an array of street canyons can vary by as much as a factor 2 under the combined effect of solar incidence and street geometry. Due to the presence of more clouds and particles, the long wave radiations emitted downwards by the atmosphere are greater over the city than over the country-side, by up to 5%. According to the Stefan-Boltzmann law (the long wave radiations emitted by a surface are proportional to its temperature to the 4th power), the radiative emission of urban grounds and surfaces is considered to be much larger than that of rural surfaces and it presents a maximum around mid-day. This flux can be as much as 20% higher over cities than over the country-side (Oke, 1979). The lower atmosphere absorption of the long wave radiation emitted by urban surfaces is also enhanced by the presence of particles and greenhouse gases like carbon dioxide, and the two effects combine to increase the temperature of the urban atmosphere. Assessment of the various terms in the radiative budget of the urban atmosphere compared to the neighbouring rural atmosphere shows that the decrease in solar radiation is quite well balanced by the effect of the lower albedos. The long wave radiations are greater in both directions and these increases tend to balance each other. Although the urban structures do transform the various terms of the radiative budget (sometimes largely), the net budget imbalances are, on average, of the same order of magnitude over urban and rural areas. Oke (1979) even considers that there is a slight deficit in the urban radiative budget compared to the rural one. Therefore, it does not seem that the radiative budget is the main direct cause of the heat island. What do we know about the urban heat budget? The sensible heat flux H is mainly due to the transfer of heat from surfaces to the air due to temperature differences between air and surfaces. The latent heat flux LE is mainly due to the release of water vapour by uncovered grounds, vegetation and the porous surfaces. But anthropogenic sources of sensible and latent heat (house and factory chimneys, traffic emissions etc.) are also present in the urban atmosphere. The magnitude of these sources is a function of the city energy consumption that results from the population density and of socio-economical factors as the average standard of life and the city organization. Anthropogenic fluxes increase 62
Ecology of the Sofia City CyanMagentaYellowBlack Page 63
Urban Climate and Air Quality - a Review
with density and with standard of living; in large, dense and rich cities, F can be greater than the solar radiative flux. In general, H is higher in cities but LE is lower. There is less vegetation and less bare soil that are potential surface water storages, but the anthropogenic sources of latent heat can be very intense and are mainly isolated, large sources (factory exhausts) that generate strong heterogeneities into the surface patterns and the atmosphere. Conduction G is much larger in cities due to the capacity of urban materials to store heat and because of the absence of vegetation. Heat storage varies between 15 and 30% of the total heat budget due to the radiative trap effect. The reduction of vegetation cover of a great number of the vertical walls increase the area collecting solar radiative flux. Although the net radiative budget has about the same value over rural and urban areas, the multiple reflections of radiation between the vertical surfaces multiply the transfers from radiative energy to heat that is stored in the city components. Heat storage accumulates during the day and is progressively released to the air during night: conduction (G) is often the largest sink of heat by day and always the largest source at night, when it almost balances the radiative heat loss. This explains the more intensive heat island during night and its maximum often observed between 3 and 5 hours after sunset: the country-side stores little heat that can be rapidly re-emitted and cools rapidly after sunset while the heat stored in and re-emitted by the artificial ground and wall materials in the city reduces the cooling of the air included in the canopy layer (Fig. 3). The parameters of the urban heat island depend also on water vapour and latent heat budgets. Water that is stored in the superficial layers of the natural ground also represents a negative latent 20
a W.m 2
15
R
10
H
G
0
F
JF
5
G 0
4
6
20
12 Solar time (hr)
16
20
24
Temparature, °C
b 15 Urban 10
Rural
5 0
0
4
6
12
16
20
24
Figure 3. Typical diurnal evolution of (a) the heat budget terms over the urban canopy (the symbols are in the text) and (b) the urban and rural temperatures (adapted by Mestayer & Anquetin, 1995 from several authors).
63
Ecology of the Sofia City CyanMagentaYellowBlack Page 64
Vassil Andreev
heat storage that is released when the water is released to the air by the vegetation. Evaporation induces a large consumption of heat in the air, about 2 500 J per gram of evaporated water (580 Cal/g). In cities, most of the water that is deposed to the surfaces is collected by the draining network to the sewers and finally to rivers and the sea. This is why evaporation is much reduced in the urban atmosphere and the latent heat that is not pumped from the energy budget is one of the major causes of the heat island. Finally, it must be noted that a negative temperature difference ∆Tu-r is sometime observed. However, while it is rarely observed in city centers, it occurs quite often in suburban areas due to the presence of artificial ground surfaces and the absence of strong heat sources and heat storages (Myrup et al., 1993). This increases the complexity of the local flows over the city and the urban “ground” description for dispersion modeling. Urban areas can induce some meso-scale perturbations in the atmosphere. One of these, demonstrated by the urban “plume” model of Fig. 1a, shows that the climatic perturbations generated by the city are transported by the wind, like a meso-scale “plume” of anthropogenically-modified and polluted air as large as the city itself. This plume can generate a series of perturbations over the downwind regions at distances of several hundreds of kilometers, as, e.g., the “Ozone plumes” during photo-oxidant episodes. Analyzing climatic records, Landsberg (1970) and Huff & Changnon (1973) showed that urban areas influenced warm season rainfall. In humid continental climate areas, urban influences on convective rainfall were most evident when the atmosphere was highly unstable and producing heavy (>25mm) rainfall events. The urban influences on cloud and precipitation processes have been difficult to measure and became the focus of several studies in the USA during the 1960’s. Changes in clouds and rainfall were suggested in an area extending 50 km east of Chicago, focusing attention on the question of the reality of urban influences on precipitation (Changnon, 1980). Studies on climate records at several North American cities (Huff & Changnon, 1973) revealed the presence of urban anomalies in precipitation and storm conditions for cities with populations greater than a million. In general, the area extent and magnitude of the precipitation anomaly were found to be related to the size of the urban area. The urban-industrial influences on the atmosphere of the St. Louis metropolitan area, and especially the regional influences on clouds and precipitation, have been studied in the METROMEX (Changnon et al., 1981). Urban causes of altered clouds and precipitation included: (a) changed incloud processes by urban-additions of cloud condensation nuclei; (b) altered fluxes of sensible and latent heat from the metropolitan area, which helped create and enhance vertical air motion; (c) increased convergence and updrafts from momentum shifts due to the city and its structures, and from differential heating, all of which affect the urban boundary layer and perturb the atmosphere; (d) enhanced moisture, largely from industrial sources. METROMEX established the causes of urban-induced cloud changes, the reality and magnitude of warm-season rainfall increases, and the types of increases in storm activity such as more thunderstorms, lightning, hail and damaging surface winds (Changnon et al., 1981). These climate changes occurred in a broad region embracing the city and a large fan-shaped area extending as far as 65 km east of St. Louis. In a study of summer rainfall over Washington it was found that the urban area caused showers, i.e. the heat island led to preferential meso-scale forcing and localized cloud development and enhancement. The sizeable changes in local, urban and regional climates due to anthropogenic activities have produced many changes in the biosphere and a mixture of winners and losers in the socio64
Ecology of the Sofia City CyanMagentaYellowBlack Page 65
Urban Climate and Air Quality - a Review
economic context (see Mestayer & Anquetin, 1995). In the lee side of St. Louis, where the climate had been altered, the 25% increase in average summer precipitation and reduced visibility due to pollutants have led to more automobile accidents, less leisure time, increased storm damages, more suburban and farm land flooding, and more soil erosion, but a net increase in crop production and increased cropland values. The migration away from urban centers leading to suburban development over the last 100 years was partly due to the less-than-pleasant climate, including poor air quality and worse human comfort in the urban center. Studies of the relocation process of wealthier urban residents and the stay-in-place decisions of affected farmers located beyond the cities is important. These findings are important for those concerned with the environmental impacts and socio-economic adjustments to large-scale changed climatic conditions. The alteration of one of the basic abiotic factors of the city ecosystem – the climate, as an effect of the process of city development, leads to a change in the degree of comfortableness of the climatic conditions for the city population. As was stated before, air pollution is also a part of the conditions defining the quality of humans in urban areas.
URBAN AIR QUALITY Urban air quality is important because of the highly urbanized human population. Many urban activities (e.g., traffic, combustion processes, industrial production) are accompanied by emissions into the air, yielding elevated concentrations of pollutants. This is especially significant when a large number of activities are concentrated together, as in an urbanized area and in big towns. Air pollution at the urban scale is the source of a range of problems: health risks mostly associated with inhalation of gases and particles, accelerated deterioration of building materials, damage to historical monuments and buildings, and damage to vegetation within and near the cities. In order for these problems to become smaller, it is necessary to understand: what characterizes urban air quality; what are the sources of urban air pollutants and the conditions of air pollution levels, and which exposures are associated with the high concentrations of air pollutants occurring in cities. Air pollution in any region is determined by two major factors – the parameters of the sources of harmful emissions and the meteorological conditions at regional and local scales. These factors determine the possible chemical transformations that could be realized in the atmosphere. The actual occurrence, frequency and level of increased urban air pollution concentrations depends primarily on the magnitude, business hours and space distribution of emission sources, on local topography (e.g., flat terrain, basin or valley) and on local meteorological conditions (e.g., wind characteristics, frequency of calm weather, occurrence of inversion layers). The fate of air pollutants is determined mainly by the height of release and the prevailing weather. Therefore, the spatial or temporal distribution patterns and effects of air pollution will range from local (up to a few tens of kilometers; time-scale less than one day), through regional (up to a few hundred kilometers; time-scale between a day and a week), to continental and global (up to a few thousand kilometers; time-scale longer than a month). Concentrations of air pollutants will vary greatly with time (daily, weekly and seasonally) and in space. In fact, air pollution at a certain location in an urban area is a result of (a) the regional background air pollution; (b) the urban background air pollution; and (c) the local air pollution. 65
Ecology of the Sofia City CyanMagentaYellowBlack Page 66
Vassil Andreev
Regional background pollution is transported over long distances, e.g. from other urban areas or by recirculation in the region. The air pollutants have had time to react with other pollutants or natural components of the atmosphere. It is difficult to estimate this pollution and it depends on climatic conditions. In the northern part of Europe, the dispersion of pollutants is mainly determined by air flow due to the prevailing winds and surface roughness. In the southern parts of Europe the dispersion is highly determined by thermal convection and local wind flow systems. Regional air pollution can be measured at stations outside greater urban areas, typically 25-50 km from the city center or by tracing the air parcels. In fact, at large scales, the dispersion in South Europe, is more complicated due to the complex flow structure. Long residence times may also lead to more complicated chemical processes, which are not easy to describe. In these cases the models for dispersion and chemical reactions need further development and our present knowledge is mostly based on measurements (Palmgren, 1996). Urban background pollution consist of the regional background pollution plus the air pollution from the urban area. It is a result of the city as a whole and the chemical reactions, which have reached a certain level and have been influenced by solar insolation and other components in the air. This pollution can be measured by a few measurement stations, located far away from local sources, e.g. above the roof tops. The representativeness of such stations for the whole urban area must be documented. Local air pollution is emitted close to the measurement/exposure location. That is why, it depends much on local sources, e.g., traffic in the street and local dispersion conditions, e.g., the surrounding buildings. Only the fastest chemical reactions are important for the locally emitted pollutants. Its concentrations are mainly determined by the dispersion. Therefore, urban pollution is the result of a superposition from large scales, regional and local emissions from natural and man-made sources, transported and dispersed by meteorological processes of the same scales, and transformed by chemical reactions in the atmosphere. Three main types of air pollution sources exist in urban areas: large stationary sources with one or a few tall stacks (point sources), small stationary sources distributed in an area (area sources) and mobile sources, i.e. traffic (line sources). The significance of any pollution depends ultimately on the type of pollutants, the resulting exposure and the health and other effects associated with this exposure. Air Quality Guidelines (AQGs) from the World Health Organization (WHO) have been taken as reference values to assess where ambient concentrations may possibly cause effects on human health and where further studies may be necessary (see, e.g., Europe’s Environment, 1995). Pollutants, including noxious substances, in the urban environment may be “primary”, i.e. emitted directly from car traffic, domestic heating, industrial or other man-made and natural processes, but may also be “secondary”, i.e. the result of chemical reactions (photochemical and non-photochemical) in the atmosphere. A specific feature of air pollution, especially in cities, is the existence of short-lasting, comparatively high concentrations, called “episodes”. They may last from one to several days, and should be discerned from “momentary” peaks of several hours’ duration, and from the “background” variability of monthly and seasonal characters. The episodes are connected with a combination of meteorological factors – poor ventilation or, in some cases, unfavorable transport and transformation processes. The episodes are often divided into “winter episodes” (“sulphuric acid smog epi66
Ecology of the Sofia City CyanMagentaYellowBlack Page 67
Urban Climate and Air Quality - a Review
sodes”), connected mainly with an increase of SO2, and suspended particulate matter (SPM) concentrations, and “summer episodes” (“photochemical smog episodes”), connected with the production of ozone (O3). Air pollution episodes usually represent local problems and should therefore be studied at an appropriate scale (horizontal resolution of about 1km) (Moussiopoulos, 1996). A requirement for such studies is the availability of a comprehensive emissions inventory including both anthropogenic and biogenic emissions. Winter episodes occur when low winds and a strong temperature inversion impede the vertical mixing and dilution of pollutants in the lowest atmospheric layers. Low temperatures increase the demand for energy, resulting in increased emissions and the further accumulation of pollutants. During these episodes, emissions related to domestic heating can be up to 70% higher than the winter season average. As well as SO2 and SPM, the winter smog mixture contains compounds such as carbon monoxide (CO), nitric acid (HNO3), nitrogen dioxide (NO2) and aerosol particles of variable content such as soot, sulphate, nitrate, ammonium, metals and organic compounds, such as polycyclic aromatic hydrocarbons (PAH). The concentrations of strong oxidants such as ozone are low during winter smog. Under winter smog conditions, daily average SO2 concentrations in rural areas can reach 400 µg/m3; it may be higher in cities (see Europe’s Environment, 1995). The most severe smog episode ever reported was in London in December 1952 when SO2 and SPM daily average concentrations reached values of about 5000 µg/m3 each. In the two-week period during and immediately after this smog episode, a total of approximately 4000 excess deaths were observed compared to such period in previous years. Europe’s Environment (1995) announced that in 61 of 87 surveyed cities of Europe, the winter smog concentration of SO2 and/or SPM exceed Air Quality Guidelines (AQGs) of the WHO. Such excesses occur in general over a large part of the city during periods (days) with poor dispersion conditions. In 24 of the 87 cities (about 28%), concentrations around or higher than twice the AQG have been measured during recent years. Sofia is among these cities. It is essential to note that the winter smog episodes often extend across several European countries due to long-range transport across the continent. Photochemical smog episodes occur under conditions of high insolation and low winds, when the concentrations of ozone and photochemical oxidants can build up to high levels which are damaging both to human health and to vegetation. Usually such situations take several hours to develop. Photochemical smog is formed particularly in land basins, and in coastal areas where air pollutants may be contained in a land-sea breeze circulation system as in Barcelona or Athens. However, under high-pressure conditions, ozone concentrations often reach serious levels over much larger areas in Europe and may reach Scandinavia and other more remote areas following transport of ozone and its precursors from the more central source areas. Because of the strong link to particular meteorological situations, the occurrence of “ozone episodes” in Europe is very variable from year to year in both space and time. The main cause of photochemical oxidant and ozone formation are the emissions of NOx (NO+NO2) and VOCs (Volatile Organic Compounds) from motor vehicles, which are densely concentrated in city areas as well as from industry (in particular, combustion of fossil fuels, solvent use and evaporation losses). VOCs present in the air from anthropogenic and natural emissions are oxidized in the atmosphere in the presence of sunlight. The process generally starts through reaction with the very reactive hydroxyl (OH) radical, by reaction with ozone or by direct 67
Ecology of the Sofia City CyanMagentaYellowBlack Page 68
Vassil Andreev
action of sunlight. The oxidation processes are complex, and NOx play an important role as necessary catalysts. A host of secondary products is formed during this oxidation; they are usually called photo-oxidants. Ozone is by far the most important of these in terms of the adverse effects it has on human health and ecosystems. Several international bodies such as WHO, UNECE (UN Economic Commission for Europe) and EC (European Community) proposed, or set guideline values for ozone for the protection of human health and/or vegetation. These guideline values indicate levels combined with exposure times during which no adverse effects are currently expected. For example, the 150 µg/m3 hourly concentration level and 100 µg/m3 for 8 hours are proposed for the protection of human health (for ozone, 1 ppb ≈ 2 µg/m3). Bulgarian guideline values for ozone are: 110 µg/m3 – average value for 8 hours – for human health protection and 200 µg/m3 (hourly averaged) – for the protection of vegetation. The hourly averaged values of 180 and 360 µg/m3 are the guidelines for information or alert for the inhabitants, respectively (see Regulation No 8, 1999). Progress in scientific understanding may lead to revision of these guideline values. The photochemical smog episodes are not only a local scale phenomenon. Oxidation takes place at time-scales ranging from a few hours to several months. During this period, the polluted air is transfered due to long-range transport. This is why the influence of these episodes on air quality and deposition of pollutants can extend over hundreds or several thousands of kilometers, i.e. they are as well a regional or continental scale phenomenon. The term “hot spot” pollution may be used to describe the high short-time pollution concentrations to which the population may be exposed when located close to pollution sources. “Hot spot” pollution includes urban streets with busy car traffic, and the pollution impact from industrial stacks in cities. In all European cities, smog occurrences and long-term average concentrations of harmful compounds such as lead, benzene, SPM and benzo(a)pyrene (BaP) receive a significant contribution from emissions from road transport. Benzene and BaP are indicators of organic compounds associated with volatile organic hydrocarbons and with soot and PAHs from small-scale combustion processes, respectively. Road transport contributes, on average, more than half to the nitrogen oxides and about 35% of the VOC emissions in Europe. While NOx emissions from stationary sources have decreased in many urban areas, emissions from motor vehicles have increased because of the growth in the number of vehicles, and the distances traveled per vehicle have been much larger than the reduction of emissions. Since the lifetime of vehicles is typically more than 10 years and in some countries as Bulgaria more than 10-15 years, reductions of emissions will be slow even if vehicle numbers and distances traveled size decrease. More rapid improvements in air quality in the short term can only be achieved by limitations on road transport in urban areas. In some cities, air pollution from road transport and domestic heating is supplemented by that from local industry. The impacted area is usually a few kilometers from the emission sources, depending on the height of the stacks and the prevailing wind directions. In many cities, monitoring stations are operated in areas where principal industries are expected to have a large impact. In many cities, these “hot spot” stations indicate a significant impact from industrial emissions, i.e., air concentrations of SO2 and/or PM are considerably higher than in the city center. Such distributions of pollutant concentrations is often observed in some industrial zones of Sofia. Most of the primary pollutants from cars, domestic heating and some industrial sources are 68
Ecology of the Sofia City CyanMagentaYellowBlack Page 69
Urban Climate and Air Quality - a Review
emitted and transformed by the chemical reactions at ground level where severe pollution levels might be found. This is why, for the preservation of a favorable air quality in urban areas, anthropogenic emissions must be controlled. Successful measures for reducing air pollution in cities over the past 30 to 40 years have included the regulation of fuel for domestic heating (low sulphur content in oil and coal, introduction of natural gas instead of coal and oil), development of district heating, reduced lead content in petrol and restriction on the use of cars in the city centers. Furthermore, dust removal from industrial sources and large boilers has resulted in major improvements. These measures have been particularly successful in reducing SO2 particles and lead emissions, and hence daily and annual average urban atmospheric concentrations, by significant amounts in many cities. However, there has been little or no evidence of similar downward trends in NOx concentrations. During smog conditions, emergency actions are taken by some cities (e.g., Rome, Milan, Athens, Prague) involving, for example, restrictions on car traffic and fuel substitution in power plants/ industries. Cost-effective energy conservation measures and substitution of fossil fuels with renewable energy resources have a large potential for reduction of emissions. As a result of such simple premises, urban air pollution management requires knowledge of the parameters related to four important aspects: a) air quality monitoring network, which provides measurements and data relevant to several pollutants, their occurrence in time, their spatial distribution and, eventually, episodes exceeding the quality standards; b) meteorological data not only from the surface but also from boundary layer measurements in the meteorological network; c) emission inventories, which provide information on the emitting source and their temporal evolution; d) numerical models, which provide information on the prediction of air quality from emission and meteorological data. These components are interrelated – monitoring data are used to validate models; meteorological and inventory data are used as input for models, which are able to provide numerical prediction. The expected meteorological situation and results from the model application and inventory data might suggest the formulation of proper control strategies. This is the most important problem for air pollution management. Air quality and the state of the atmosphere are usually fixed through corresponding systems for air quality monitoring and for meteorological measurements. They have the most essential module - the air quality monitoring and meteorological networks, which are made up of many measuring stations. One of the main purposes of these systems is to produce information for understanding the basic processes, which are responsible for atmospheric pollution and to take measures required for reducing the risk for health and the environment from air pollution. The air quality monitoring stations are composed of many chemical and physical-chemical sensors, which provide the concentration of selected compounds. Unfortunately, only a few parameters can be measured automatically and, if a complete picture of atmospheric processes is needed, manual chemical procedures must be employed. These are required for inorganic as well as organic components. Two kinds of permanent networks and systems for air quality monitoring are used in dependence on the necessity to take either long-term or short-term measures for reducing the risk from pollution. They are : networks (systems) for general monitoring of air quality, and air pollution informativewarning systems (networks). General monitoring systems are considered the basic systems in urban areas. Manual measurements of 24-hour concentrations, with samples taken on a site and then 69
Ecology of the Sofia City CyanMagentaYellowBlack Page 70
Vassil Andreev
analyzed in a laboratory, still predominate in countries of Central and Eastern Europe. Data produced in the system are used for long-term environmental programs; for assessment of the effectiveness of a long-term policy, of the air pollution level in connection with the national ambient air quality standards, of the trends and validation of mathematical models; for documentation of population exposure to air pollution; for town and country planning; for scientific investigations etc. The stations of the urban general monitoring systems are located in inhabited areas, where the expected long-term concentrations of a particular pollutant are higher than in other areas of the town as well as, where concentrations of a particular pollutant may be treated as average for densely populated parts of the town. These stations can also provide feedback for many other specific purposes. The informative-warning systems are also of two kinds – warning systems, based on air-pollution forecast, which, in most cases are for information only, and alarm systems, based on actual measured concentration values, which are connected with enforcement of compulsory preventive actions. In the case of alarms, the responsibility of an informative system is higher, because the operation of emergency emission-reduction system is expensive and may be troublesome for citizens (e.g., traffic limitations). The informative-warning systems collect data in real time by means of continuous monitors. The main task of these systems is to provide data for short-term pollution abatement procedures. They should be established only in densely populated areas with frequent occurrence of high concentrations of pollutants caused by different sources and especially from a particular industrial plant or from distant plants as well as from traffic. The meteorological data, necessary for the functioning of the informative-warning systems, are received through meteorological observations either at the same point where air pollution is measured, or at the nearby meteorological station. During field experiments these two types of data – for the concentrations of pollutants and atmosphere physical parameters, are most frequently conducted simultaneously at the same points. It can be noted that during episodes extraordinary measures can be applied to protect the population. There are two kinds: active measures – emergency reduction of local emissions, and passive measures for a reduction of health hazards. The procedures for the realization of emission inventories are well established. But the assessment of the emissions, due to the trans-boundary transport of pollutants, represents not only a national, but also an international problem. This fact necessitates the establishment of an inventory not only of pollution sources in separate plants, but also integrated inventories of emissions for a certain city, area, country, region and even a continent. The first inventory system of this type in Europe was established in the 1980s and was called CORIN (CO-oRdination of INformation on the environment), and one of its projects focused on air, and was called CORINAIR. In the inventory procedure, conducted in 1990, around 250 activities and 8 pollutants were covered for Europe in accordance with the development of continental modelling within the EMEP programme (Cooperative Programme for Monitoring and Evaluation of the Long Range Transmission of Air Pollutants in Europe). The inventory methods were improved in 1995 – CORINAIR’95. Another set of methods of wide international use is the IPCC/OECD (Intergovernmental Panel of Climate Change/Organization for Economic Cooperation and Development). It is related to the inventorying of basic “greenhouse gasses” and also to the assessments of gasses influencing stratospheric ozone. 70
Ecology of the Sofia City CyanMagentaYellowBlack Page 71
Urban Climate and Air Quality - a Review
The processes that determine the distribution of gas pollutants in air and that must be considered in the atmospheric dispersion models are: advection and convection (horizontal and vertical transport); diffusion on the three directions in space; dry deposition in the moment of contact of pollutants with the earth surface; the washing off of pollutants by precipitation (wet deposition); chemical transformations and/or radioactive decay. When having pollutants, carried by aerosol particles, one must additionally include the processes of gravitational sedimentation and transformation of particles. These depend on the spectrum of particle sizes that itself changes in time as a result of interactions among particles in a gas or liquid environment (processes of decay or coagulation). These processes, as well as the chemical transformations, develop differently in the dry atmosphere, in a cloud environment and in precipitation. Adding that because of chemical transformations some gasses transform into aerosol particles and vice versa we can obtain a notion of the complexity of the distribution of air pollutants. Each of the latter processes can be described mathematically with a different degree of complexity and with different time and space characteristics (scale and resolution). It is obvious that with such a number of processes in a continuous medium like the atmosphere, the only way to describe them is through numerical modelling. In the international literature there is a large number of numerical dispersion models with different purposes and degrees of complexity. The different models can describe the distribution of one or many pollutant, emitted in the atmosphere by different types of sources at various time and space scales. In accordance with the approach for the description of dynamic processes, the models can be separated generally into 4 classes: Gaussian plume models; Lagrangian-Eulerian cloud models (puff models); Eulerian models; Lagrangian particle models (Monte-Carlo models). Plume models are the simplest and historically the first models, based on many simplifications of the dispersion processes. They describe the distribution of a weightless gas from a continuous point source in an environment of constant wind speed (stationary conditions). The result of this idealized setting is the formation of a plume that widens while drifting away from the source under the influence of air turbulence. The distribution of the pollutant concentrations in the plume follows a Gaussian process. During the development of studies and increased requirements of practice, the Gaussian models have experienced a considerable evolution, a result also of the gathering of much experimental material. First, the improvements encompass the means of reading the intensity of the turbulence, which is a result of the atmospheric boundary layer stability. The most utilized description is through the Pasquil classes where, on the basis of ground measurements of different air characteristics, a gradation of the turbulence intensity is entered, and from there – a gradation of the different parameters of the model as well. At first, this affects the values of the dispersion coefficients in the Gaussian formula – a quicker or slower widening of the plume depending on stability. Other parameters of the model also depend on the Pasquil classes such as the wind profile with altitude, the degree of the plume meandering, etc. Some considerable improvements of this type of models were achieved with the implementation of the so-called plume-rise, i.e., the elevation of the plume axis over the earth surface as a result of overheating and the forced speed of emission of the pollutants in the stack. Other improvements allowed the application of the Gaussian model with line, area and volume sources, for descriptions of aerosol pollutants or the flows around the buildings. An adequate parameterization of dry and wet deposition rendered the 71
Ecology of the Sofia City CyanMagentaYellowBlack Page 72
Vassil Andreev
possibility to calculate, apart from concentration, the deposition of the pollutants on the earth surface. These type of models are widely applied in modelling the distribution of air pollutants in urbanized areas. These models are also used for regulatory purposes, as, for instance, determining the suitable height of stacks of new-built plants and assessment of the impact of existing and future sources on the environment. In the latter, case series of hourly meteorological measurements for a long period of time are used for meteorological input of models, as well as climatic characteristics of the area like the complex rise of wind (direction-speedstability class). In spite of their considerable improvements, Gaussian models have essential limitations. The greatest one is that they can only be applied to limited areas of up to 30 km of homogeneous surfaces, and for time scales not greater than 1 hour, for which it can be assumed that no substantial changes in the wind field occur and the model can only operate with a single wind value. The puff models are similar to the Gaussian models, but here the limitation for stationary and homogeneity of the wind field no longer exist. Therefore these models can be applied at various time and space scales. The emissions from sources with a single action in this type of models are approximated with a puff of polluted air that is caught and carried by the air flow while simultaneously the puff widens under the influence of air diffusion. The continuous sources are approximated with a series of puffs emitted at some time intervals. It is accepted that there is Gaussian distribution of the substance in the puffs. As long as the diffusion equation is linear, the principle of superposition allows for the calculation of the total field of concentration by a summation of the pollution fields created by the separate puffs. The principle of superposition allows for the description of pollutants at which linear transformations occur. Examples include simple schemes for sulphur and nitrogen. More complex schemes such as photo-oxydation are generally not linear and cannot be used in puff models. The puff models are most suitable for a small number of point sources and especially for a single point source as is approximated, for example, a nuclear accident. Eulerian models are based on the Euler approach for describing hydrodynamic fields. Here the observer is positioned at a fixed point in the fluid and the alteration of the characteristics at this point is a result of the passage of all new particles, carrying different characteristics. In the Eulerian models the half-empirical advection-diffusion equation is solved (see e.g. Berlyand, 1985). The space within these models is discrete and that is why the input information and the calculated characteristics are related to these points, i.e. these are values characteristic for the cells around each point. Due to this, the resolution capacity is an essential characteristic of the models and of the input information. The processes with a smaller scale than the step of the network are parameterized by expressing their characteristics through the parameters of the processes having scales encompassed by the model, i.e. with scales larger or equal to the step. Different numerical schemes can be used in modelling the separate terms in the advectiondiffusion equation and that determines the variety of models, not only according to the type of pollutants and the task scale, but also the memory and calculation rate of the present equipment, the capacity of the interface for feed-up with information, etc. The Eulerian models are the most developed means for dispersion calculations (see Syrakov & Ganev, 1994; Syrakov, 1997). These are the most complex existing models, due to their 72
Ecology of the Sofia City CyanMagentaYellowBlack Page 73
Urban Climate and Air Quality - a Review
capacity to accommodate complex, not-linear chemical schemes with around 100-200 air pollutants. Due to the great number of points in the model network, the number of network values can reach a few million. This places high demands for computer memory and calculation rate. This is why Eulerian models are implemented on the most powerful supercomputers with vector and parallel calculation. These are usually coupled with the most complex global and regional atmospheric models providing the necessary input information for the fields of the meteorological elements. Lagrangian models developed more recently. The name comes from another approach for describing hydrodynamic quantities – the Lagrangian. Here the observer is positioned in a fixed fluid particle and moving with it, he follows the alteration of its characteristics. The Lagrangian models describe most directly the transport and diffusion in the atmosphere. The pollution source emits a great number of particles and each one is observed. For one step in time the particle is moving in accordance with the values of the wind components in the point where it is situated. Along with this, some shifting is necessary as a result of turbulent diffusion. Shifting is calculated from three coordinates and its value is determined using a random numbers generator (the Monte Carlo method). Maximum shifting depends on the turbulent characteristics of the atmosphere, the stability of layers, etc. Thus, each particle, apart from moving in the flow, receives pushing away and in result the group (ensemble) of particles forms a constantly widening puff. The Lagrangian models are suitable for calculating the distribution of emissions by single momentary sources, but require great memory and calculation rates. Still, great difficulties exist in describing the other processes such as chemical transformations, dry and wet deposition, etc. Nevertheless, due to the direct approach for describing the two most important processes, these models are actively developed and used. Recently, air pollution models have reached a sufficient degree of accuracy. The simulation of dispersion of pollutants over urban areas is the best performed using meso-scale models. Dispersion modeling in the canopy layer is of primary importance for human life and activity. The structure and stratification (dynamic and thermal) of the air flows over cities are important for any kind of model for the dispersion of pollutants over this area. The simplest approach in the simulation of dispersion over meso-scale urban areas, is that flow over an urban area is similar to the flow over a “normal” rough surface, with a given, large roughness length zo and a defined surface heat flux. Assuming that the city is large enough to ensure a locally “sufficient” horizontal homogeneity, the Monin-Obukhov’s similarity theory can be apply to meso-scale dispersion calculations (see Mestayer & Anquetin, 1995), although their applicability to cities seems highly questionable. Then the Monin-Obukhov parameters can be defined in the usual way and flow structure can be characterized by the surface stress ru*2 , and sensible (Hs) and latent (LE) heat fluxes. Due to the great height of the roughness elements, it is needed to introduce a zero-plane displacement height d in the parameterization of the mean wind profile. zo and d can be either computed from measured profiles of U(z) or deduced from empirical relationships, relating these variables to the density and geometric characteristics of the roughness elements (buildings, trees etc.). These calculations can be of value when the details of the ground structure cannot greatly influence transport and diffusion, as e.g. buoyant plumes emitted from outside the city and transported high above ground level in convective meteorological conditions. 73
Ecology of the Sofia City CyanMagentaYellowBlack Page 74
Vassil Andreev
The most recent investigations of flow and dispersion in urban and industrial areas are represented in the handbook of Hanna et al. (2001), which is summarized in the paper of Hanna and Britter (2001). A review of the literature was carried out, which led to the recommendation of a hierarchy of methods for estimating the parameters zo (a roughness length), and d (a zero-plane displacement height) that are introduced in the parameterization of the mean wind profile. The simplest methods, based on land-use classes, lead to estimates of zo of about 1 m and d of about 5 m in urban and industrial areas. More accurate estimates of zo and d have been suggested by many researchers, based on parameterizations involving the height of the roughness obstacles, Hr, and measures of their spacing, λf (λf=obstacle frontal area/obstacle lot area). Hr and λf can be estimated based on experience of certain well-known urban and industrial land-use types (Grimmond and Oke, 1999), or can be calculated in detail using data on actual building shapes and sizes in an area. According to Hanna and Britter (2001), the average building or obstacle height, Hr is about 10 m and the values of the obstacle densities, λf are between 0,1 and 0,3. The methods for determining zo and d at different values of Hr and λf, as well as formulas for their calculation are presented in Hanna and Britter (2001). Given the estimates of zo/ Hr and d/ Hr using these methods and assuming neutral conditions, the friction velocity can be estimated from a wind speed observation or estimate, u, at some height, z, which is above Hr: u* = 0,4 ln
u (z ) ) (z–d )
(8)
z0
Hanna et al. (2001) show that within the urban or industrial site, the wind speed, uc, below the obstacle heights, Hr, is given by: uc/u*=(zo/2 Hr )-1/2
or (λf ) -1/2
(9)
The log-linear wind equation (1) is assumed to apply down to a height, zint , such that u(zint)=uc. The turbulent velocities in urban or industrial regions are given by the following relations at all heights, above and below Hr (Roth, 2000):
σu/u*=2,4,
σv /u*=1,9,
σw/u*=1,3
(10)
Given zo, d, Hr and lf and using standard Monin-Obukhov similarity theory it is possible to express wind and turbulence profiles as a function of height, for elevations near the ground well bellow Hr to elevations at 10 or more times Hr. These fundamental boundary layer formulas are valid over a very rough surface, as long as Hr is less than the surface boundary layer height (about 50 to 100 m). The friction velocity, u*, is a key scaling parameter for all boundary layer phenomena. The effects of roughness on stability (as measured by Monin-Obukhov length, L) can be estimated, using the fact that an increase in u* over a rough surface can lead to much larger L (due to the dependence of L on u*3). Transport and dispersion models for air pollutants can typically be run in the hypothetical or “what-if ” mode, or in the real-time or retrospective mode. In the first mode, the worst-case conditions may be arbitrarily chosen as light winds and stable conditions. But in the second mode, where observations must be used, the available observations are usually not from the best location. 74
Ecology of the Sofia City CyanMagentaYellowBlack Page 75
Urban Climate and Air Quality - a Review
Hanna and Britter (2001) argue that the wind speed at a height of 30 m, say, will be similar at the site of interest (with z02) and at a flat measuring station nearby. Then U2 (10m)=u1(30m).ln(10m/z02)/ln(30m/z02)
(11)
Because of the increased mechanical turbulence over an urban or industrial site, at any given time the stability is closer to neutral than the stability over the airport or other nearby flat site. This is because the increase in z0 over the rough site causes increases in u*. And since the stability parameters such as L depend on u*3, a two-time increase in u* results in an eight-time increase in L. A less simplistic approach considers the atmospheric flow transformation due to the internal boundary layer (IBL) in development along the wind direction as 2-D models (Figs. 1 and 2) as well as in formula (4), where several empirical parameterizations of the IBL depth zIBL as a function of the fetch x are shown. Melas & Kambezidis (1992), reviewing some recent models of the thermal internal boundary layer (TIBL), divide them in 3 categories: (a) empirical models of the form zIBL=a.xb; (b) similarity models where the growth of the TIBL is parameterized by analogy with the spread of a smoke plume from a ground-level source (an analogy quite valid in the case of near-neutrality for a very weak transition in roughness and surface heat flux, based on the equality of the diffusivities for momentum and matter); (c) slab models assuming that the TIBL is composed of a mixed layer, where the vertical profiles are uniform, bordered by thin layers, surface layer at the bottom, and entrainment layer at the top. In these assumed thin layers, the temperature gradients are forced to match the ground temperature and the unperturbed temperature profile above the TIBL top, identical to the profile upstream of the transition (Fig. 4). Melas a)
δ (x)
Equilibrium layer Advection
Transition layer δ1 (x) Equilibrium layer
z02
z01 θ
b)
z
zTIBL (x) layer ibrium Equil
z01 θs1
Diffusion
z02
Mixed layer
TIBL
Surface layer θs2
Figure 4. Schematic representation (adapted by Mestayer & Anquetin, 1995) of the development of an internal boundary layer at a transition to a rougher surface (z02> z01): (a) after Hunt & Simpson (1982) and Folcher (1989) (in the case when z02< z01 , the diffusion is downwards in the transition layer); (b) after Melas and Kambezidis (1992).
75
Ecology of the Sofia City CyanMagentaYellowBlack Page 76
Vassil Andreev
& Kambezidis (1992) have noted that the “slab” models as those of Gryning & Batchvarova (1990), which are of higher complexity, present good agreements, at the mesoscale, with the experimental data over the Athenian urban coastal area, when the meteorological conditions are close to neutrality and the fetch is large. Mestayer & Anquetin (1995) showed that none of these models are able to correctly represent the structure of the flow altogether at short and long fetches (the slab models are better for long fetches, mesoscale dispersion, while similarity models are well adapted to short fetches, small scale dispersion) and to represent the structure of the lowest layers of the urban areas (just over the roofs). Experimental data obtained inside large cities show that the vertical structure of the low atmosphere is more complex than that in Fig. 4. As is seen in Fig. 5, there is a perturbed layer over the canopy layer where Monin-Obukhov surface layer similarity theory does not hold. Below the roof top height, the urban canopy layer is characterized by flow recirculations and radiative trappings. Just above the roofs, the flow is close to horizontal on average but highly perturbed by individual building wakes and heat flux variations. Above this transition layer, if the city fabric is not too heterogeneous, the atmosphere eventually “feels” the lowest layers as a rough warm terrain; its structure can resemble that of a surface layer at (local) inertial equilibrium, and it can be characterized by locally-defined Monin-Obukhov similarity parameters. This “local surface layer” can be capped by a “mixed layer” that actually is itself a growing TIBL. Simulations of mesoscale dispersions of pollutants generated inside the canopy and diffused upand out-wards, or of those coming from the city outskirts and eventually penetrating the canopy down to the pedestrian level, must take into account this multiple layer structure of the urban lower atmosphere that is largely dependent on the structure of the canopy elements. The basic elements of the urban canopy are the isolated or in-group buildings and the street-canyons. The geometry of the buildings, their orientation and angles, the properties of their surfaces, their albedo and emissivity, the shadowing effects and the reduction of the optical aperture to the
a)
b)
zi
z Logarithmic layer
Transition layer Hr z0 + d d
z0
Urban Boundary Layer
U (z)
z
Mixed layer zmax = 0.1zi Inertial layer Surface layer
Roughness sublayer D
Hr
0 Figure 5. Schematic representations of the urban boundary layer presented by two groups of authors, pointed at Mestayer & Anquetin (1995), which have adapted these schemes, too.
76
Ecology of the Sofia City CyanMagentaYellowBlack Page 77
Urban Climate and Air Quality - a Review
sky are the physical factors intervening in the radiative budget and significantly modify the energy budget. In addition to a very large number of isolated experiments, a large part of our knowledge comes from the continuous series of studies and numerical simulations of the groups of the University of British Columbia (see, e.g., Nunez and Oke, 1976; Oke, 1973 to 1988) and the University of California, Los Angeles (e.g., Todhunter & Terjung, 1988). The former used mainly an empirical and experimental approach that allowed to obtain most of the basic knowledge into the processes taking place in real streets, but are difficult to extrapolate to other sites and to build quantitative predictions. The second group essentially used and developed numerical models to realize series of numerical experiments and process studies in order to assess the influence of the different climatic and anthropogenic parameters of importance in the energy budget. The North-South orientation of streets improves the maximum radiative heating of the ground at mid-day. The maximum net radiative imbalance, with a clear dry atmosphere, is on the order of 400 W/m2 at noon. About 60% of the energy is then transformed into sensible heat in the air and 30% is stored in the canyon materials; the remaining 10% are either stored in the ground materials or used for evaporation. Oke (1988) shows that the effective albedo is a function of the solar elevation angle and the canyon geometry. He expressed this dependence by means of a sky view factor that is a function of the height-to-width ratio Hi/W where Hi are the building heights and W the street width. An essential issue in the prediction of the energy budget is the effect of the multiple reflections of the incident radiations over several surfaces of the canyon (radiative trap). It goes from a billiard 3-band looking trajectory in the case of moderately high buildings to a very large number of successive reflections in the case of arrays of elevated mirror towers (see Mestayer & Anquetin, 1995). The radiative trap, that is a major cause of heat storage, can also be seen as a reduction factor of the effective albedo. It can be summarized that the present knowledge of the process’ interactions is insufficient (Todhunter & Terjung, 1988). There are three aspects of the dispersion of pollutants inside the canopy layer: dynamic, thermal and chemical. Most of the knowledge of the in-canopy dispersion comes from simulations in aerodynamic tunnels, essentially in isothermal conditions, very few dynamic measurements on site, and a handful of neutral tracer dispersion experiments at small scales in the atmosphere (see, e.g., Mestayer & Anquetin, 1995; Hunter et al., 1992; Oke, 1988). Essentially, two types of canopy flows have been simulated in wind tunnels: flows around and behind isolated buildings, and in groups of buildings. Real size measurements of flows in streets are rare. Examples include the Lagrangian balloon experiments of DePaul & Sheih (1986) in a long street of Chicago, USA; Nakamura & Oke’s (1988) measurements in one cross-section of a street in Kyoto, Japan as well as such a cross-section of Baranger (1984) in Nantes, France and measurements at the canopyatmosphere interface in Zurich of Rotach (1991). More details on the aerodynamic effects of urban structure can be found in Mestayer & Anquetin (1995). In the 1980’s, there were few models specifically developed for simulating contaminant dispersion within the urban canopy. Only some of these took into account the effects of a building’s wake on the dispersion of a plume, by extending the Gaussian plume approach or by simulating the building wake by a secondary source of contaminants. More references can be found in the controversial article of Ramsdell (1990) and the subsequent discussion of Briggs et al. (1992). It must be noted that these studies have been performed mainly in near neutral conditions, with 77
Ecology of the Sofia City CyanMagentaYellowBlack Page 78
Vassil Andreev
little or no thermal influence on the flow. The numerical simulations realized by the group in Nantes, France (see Mestayer & Anquetin, 1995) clearly demonstrated that temperatures of the street walls and the ground largely influence the in-street flow structures and velocities, and the dispersion of the contaminants. Many of the concepts that are now considered as “standards” must certainly be revisited by combining radiative, thermal and dynamic simulations. New on-site experiments with a large concentration of sensors allowing extremely detailed documentation of flows and climate in typical canopy geometries are necessary to validate the available mathematical models. Present-day dispersion models in urban and industrial areas are divided by Hanna et al. (2001) into the following 5 groups (Hr – height of the roughness obstacles): 1) Dispersion models for clouds extending above Hr. These models often require flow parameters such as z0, d and u*. For passive releases at any height, Hanna et al. (2001) suggested the Gaussian model for concentration in terms of the dispersion coefficients, which are expressed in terms of u* and (in the far-field) the integral time scale Tl. 2) Dispersion models for clouds released above or near Hr. These models are similar of the previous models. The implication of this choice is that the flow in which the release is dispersing is determined by the underlying roughness. This will be a less valid assumption for releases well above Hr and when the release is close to the upwind edge of the roughness but this would be an unlikely situation. This statement holds for passive and light gas releases. It is likely to hold for heavy gas releases until the cloud trajectory descends well inside the roughness obstacles. 3) Dispersion models for clouds below Hr. There is evidence that a Gaussian model is appropriate for this case, but there are no widely-used operational models for this scenario. They require specification of a characteristic velocity, uc, and correlations for the dispersion coefficients. For a continuous plume, the concentrations initially decrease with increasing roughness but eventually become constant and then may increase for very high roughness. For puffs, the correlations predict a monotonic decrease in concentration with increasing roughness. These statements are applicable for passive and light gas releases. For heavy gas releases, a modification would be required to existing operational models; the principal modification being the modification to the velocity profile within the roughness. 4) Dispersion models for clouds at a downwind of the roughness. These models only account for the effects of roughness downwind of the source. Thus the roughness to be considered is an average of the roughness between source and receptor. This approach may lead to concentration overprediction close to the source for the situation where the source is near the downwind edge of the roughness. 5) Dispersion models for clouds released upwind of the roughness and traveling into and through the roughness. This approach considers the roughness between source and receptor, and by implication uses an advection velocity determined by this roughness. This is a very simplistic model of the complex flow as the atmospheric boundary layer impinges upon a finite area of increased roughness. Urban planners, local decision-makers and city leaders, as well as responsible officials of the environment, need more complicated data to understand the results of their polices, especially concerning traffic management, to assess the actual air quality with regard to secondary pollutants, to finely predict all-type impacts of new chemical sources, as well as to foresee the effects of new regulations. The need for estimations of secondary pollutant concentrations is especially 78
Ecology of the Sofia City CyanMagentaYellowBlack Page 79
Urban Climate and Air Quality - a Review
demanding, since they depend on mixing and reactive processes: this implies a simultaneous knowledge of chemical kinetics and of transport-dispersion behaviour of primary and intermediate compounds. Recent research at the canopy scale aiming at more general descriptions of urban climatology are developed towards three main directions (Mestayer & Anquetin, 1995): (a) The combination of local on-site measurements with numerical simulations of the radiative heat budget allows to further develop comprehensive urban radiative models and operational codes; (b) the intensive use of 3-D computational fluid dynamic numerical codes with high resolution and a complete turbulence model, allows systematic studies of in-canopy flows and canopy-atmosphere exchanges as a function of building geometries and surface temperatures; (c) the development of street air quality models, that were originally of a box-type with primary pollutants, include now geometrical features of streets and of several chemical reactions to predict secondary pollutants. These three approaches are merged into the recent dispersion models. A very important fact in recent years is that co-ordinated efforts have been undertaken within the European framework to use and/or adapt mesoscale meteorological codes to the study of the urban climate as, e.g., the Athenian Photochemical Smog Intercomparison of Simulations (APSIS). It was shown that several mesoscale models can predict correctly the air flows at the regional scale but that practically none can detail the flows just above roof level responsible for pollutant dispersion in the complex area of Athens, an urban, coastal, hilly and mountainous site. Furthermore, serious tests of the models, especially those combining dynamics and photo-chemistry are required to organize experiments with extensive documentation of micro-meteorology, emission inventory and compound monitoring. Another important development of meso-scale studies is the comprehensive use of satellite observations to analyze the city fabrics and land uses over entire urban areas. The combinations of several channels of the same satellite (Lee, 1993) or of different instruments as, e.g., NOAA’s AVHRR, Landsat and SPOT visible and IR channels allows detailed analysis of the apparent ground temperature, albedo, reflectance and heat emission. Pixel sizes can be as small as 20 m (SPOT) and multiple image processing could even provide final resolutions smaller than 10 m. Similar airborne radiative measurements can also be of importance for calibration purposes and finer scale analysis (Takamura, 1992). In conclusion, concerning pollutant dispersion at the mesoscale, it must be noted that the Hybrid Plume Dispersion Model (HPDM) is an interesting attempt at solving the problems because it takes into account urban land use categories and their distribution along the wind path as in Fig. 2a (Hanna & Chang, 1992). Such an urban operational dispersion system (LOSTRAC) is being constructed in co-operation by French and Danish institutions (see Mestayer & Anquetin, 1995). Furthermore, reliable parameterizations of the roughnesses of urban gorund and of heat fluxes do not exists, and of the transformation of the orographic effects by constructions such as, e.g., downslope “street winds” and “quartier breezes” (Fig. 2b). Concerning urban sub-mesoscale research, it appears that the multiplication of measurements will not suffice to assess and to predict flow structure, dispersion processes and air quality or pollutant contents, and their impacts, because of the large structural, temporal and spatial variability of the urban lower atmosphere. The improvement of our knowledge on the combined physics and photochemistry of the urban atmosphere calls for the development of a new generation of 79
Ecology of the Sofia City CyanMagentaYellowBlack Page 80
Vassil Andreev
numerical simulation codes. Indeed, partial simulations of the radiative, thermal and dynamic processes taking place within the urban canopy are still essential, but the assessment of photochemical compound behaviours require combined simulations, including fast photo-chemical reactions. More impact studies at the scale of the urban area itself require sub-mesoscale models able to simulate the dynamics and the physics altogether over the whole atmospheric boundary layer height and a meso-γ scale, and at the decametric scales of the local convective movements in the lower layers. These processes are responsible for the pollutant mixings and the chemical interactions between compounds generated by sources as different as industrial stacks, intense or dispersed traffics, individual heating or remote “background pollution”. Not only the detailed prediction, but also the real time control of the air quality demand the development of these complex numerical models to understand the relations between the physical-geographical conditions, the regional meteorology, the local pollutant sources and the monitoring measurements. The development of sub-meso models will allow us to study the interactions of the city fabrics and urban sources of heat and pollutants with the surrounding orography. This is of special importance for cities located in valleys or valley systems, where the urban radiative budget interacts with the along-valley and katabatic cross-valley winds; this can produce extreme pollutant traps. This model development is possible only if the structure of the flow in the intermediate layer just above the roofs is better understood (Fig. 5). In sub-meso models, the urban canopy structure is usually replaced by a “terrain”, composed of a patchwork of homogeneous parts of a city, sometimes named “quartiers”, for which specific ‘wall laws’ are defined. These laws are more complicated than the usual roughness length laws based on relations (formula 8) from the previous text. New parameterizations are also needed to improve operational models; they include not only drag and thermal fluxes but also the ability of the canopy to exchange matter with the atmosphere downwards and upwards. Also, they will not depend exclusively on the average structure of the quartier but as well as on its position with respect to the distribution of quartiers and wind direction. In the modeling activity, the aerodynamic quartier is often defined as a part of a city that has homogeneous features for numerical flow and dispersion simulations. To develop and assess these parameterizations, three types of research are urgently needed, preferably in co-ordination: (a) numerical simulations of the development of multiple three-dimensional boundary layers over patchy terrains; (b) numerical simulations of the vertical exchanges in the interfacial layers between canopy and atmosphere, combining dynamic and thermodynamic models in homogeneous quartiers; (c) on-site experiments, documenting with great detail, at the same time, the flows at street level and the flows in the atmospheric layers from the roof level to the inversion height.
REFERENCES Andreev V., Ch. Branzov, E. Koleva, A. Tsenkova, J. Ivantsheva, P. Videnov. Climate and Human Comfort of Sofia. In: Penev, L., Niemelä, J., Kotze, D. J. & Chipev, N. (eds.) Ecology of the City of Sofia: Species and Communities in an Urban Environment, pp. 25-54. Sofia-Moscow: Pensoft Publishers. Andreev V., D. Syrakov, A. Tsenkova, J. Ivantsheva, P. Videnov. Air Quality of Sofia City – Meteorological Aspects. In: Penev, L., Niemelä, J., Kotze, D. J. & Chipev, N. (eds.) Ecology of the City of Sofia: Species and Communities in an Urban Environment, pp. 83-127. Sofia-Moscow: Pensoft Publishers.
80
Ecology of the Sofia City CyanMagentaYellowBlack Page 81
Urban Climate and Air Quality - a Review BC-EMEP, 1994-1999. Bulgarian contribution to EMEP. Annual Reports for 1993-1998. NIMH- Sofia & EMEP’s Meteorological Syntesizing Center-East –Moskow; pp. 93, 120, 93, 37, 42, 38. Berlyand M.1985. Air Pollution Forecasting and Regulation. Gidrometeoizdat, Leningrad. 272 pp. (in Russian). Briggs G., A. Huber, W. Snyder, R. Thompson, J. Ramsdell. 1992. Discussion on “Diffution in building wakes for ground-level releases”. Atmospheric Environ. 26B: 513-517. Changnon S. A. 1980. More on the La Porte anomaly: A review. Bull. Amer. Meteor. Soc. 61: 702-711. Changnon S., R. Semonin, A. Auer, R. Braham, J. Hales. 1981. METOMEX: A Review and Summary. Meteorol. Monogr. No118, American Met. Soc., 181 pp. DePaul F.T., C.M. Sheih. 1986. Measurements of Wind Velocity in a Street Canyon. Atmospheric Environ. 26B: 455-459. Europe’s Environment. The Dobris Assessment. (Stanners D., Ph. Bourdeau, Eds.). 1995. European Commission: DG XI and Phare. European Environment Agency. Copenhagen. 680 pp. Grimmond C. S. B., T. R. Oke. 1999. Aerodynamic Properties of Urban Areas Derived from Analysis of Surface Form. J. Appl. Meteorol. 38: 1261-1292. Gryning S. E., E. Batchvarova. 1990. Analytical model for the growth of the coastal internal boundary layer during onshore flow. Quart. J. R. Met. Soc. 116: 187-203. Hanna S., R. Britter, P. Franzese. 2001. The Effect of Roughness Obstacles on Flow and Dispersion at Industrial and Urban Sites. AlChE/CCPS, New York. 155 pp. Hanna S., R. Britter. 2001. The effect of roughness obstacles on flow and dispersion in urban and industrial areas. In: Harmonisation within atmospheric dispersion modeling for regulatory purpose. Belgirate, Italy – 28-31 May 2001, pp. 266-270. Hanna S.,J. Chang. 1992. Boundary-layer parameterization for applied dispersion modeling over urban area. Boundary-Layer Meteorology 58: 229-259. Huff F.A., S.A. Changnon. 1973. Precipitation modification by major urban areas. Bull. Amer. Meteor. Soc. 54: 1220-1232. Hunter L.J., G.T. Johnson, I.D. Watson. 1992. An Investigation of Three-Dimensional Characteristics of Flow Regimes within the Urban Canyon. Atmospheric Environ. 26B: 425-432. Jensen N. O. 1981. Studies of the atmospheric surface layer during change in surface conditions. Colloque Construire avec la vent. CSTB Nantes, Tome 1. 4.1-4.20. Kondo J., H. Yamazawa. 1986. Aerodynamic roughness over an inhomogeneous ground surface. BoundaryLayer Meteorol. 35: 387-407. Landsberg H. E. 1970. Man-made climatic changes. Sciences 170: 1265-1274. Lee H.1993. An application of NOAA AVHRR thermal data to the study of urban heat islands. Atmospheric Environ. 27B: 1-13. Melas D., H. D. Kambezidis.1992. The Depth of the Internal Boundary Layer over an Urban Area under Sea-breeze Conditions. Boundary Layer Meteorology 61: 247-264. Mestayer P., S. Anquetin. 1995. Climatology of Cities. In: Diffusion and Transport of Pollutants in Atmospheric Mesoscale Flow Fields. A. Gyr and F.S. Rys (eds). pp. 165-189 Kluwer Academic Publishers. Moussiopoulos N., Th. Flassak, Ch. Kessler. 1992. Modelling of photosmog formation in Athens. In: Air pollution modeling and its applications IX. A. Van Dop and G. Kallos (Eds.), NATO 17, Plenum Press. Moussiopoulos N. 1996. Air Pollution Modelling and its Interactions with Experimental Studies – The Situation in Southern Europe. In: Urban Air Pollution. Monitoring and Control Strategies. (I. Allegrini and F. DeSantis. Eds.). pp. 317-328. NATO ASI Series/2, Environment, v. 8.
81
Ecology of the Sofia City CyanMagentaYellowBlack Page 82
Vassil Andreev Myrup L., C. McGinn, R. Flocchini. 1993. An analysis of microclimatic variation in a suburban environment. Atmospheric Environ. 27B: 129-156. Nakamura Y., T.R. Oke. 1988. Wind, Temperature and Stability Conditions in an E-W Oriented Canyon. Atmospheric Environ. 22: 2691-2700. Nunez M., T. Oke. 1976. The energy balance of an urban canyon. J. Appl. Meteorol. 16: 11-19. Oke T.R. 1973. City size and the heat island. Atmospheric Environ. 7: 769-779. Oke T.R. 1979. Review of urban climatology 1973-1976. WMO, Technical Note No539. Oke T.R. 1981. Canyon geometry and the nocturnal heat island: Comparison of scale model and field observations. J. Climatology 1: 237-254. Oke T.R. 1988. Street Design and Urban Canopy Layer Climate. Energy and Building 11: 103-113. Oke T.R. 1990. Bibliography of urban climate 1981-1988. WCAP-15, WMO/TDNo397, 62p. Oke T., G. Johnson, D. Steyn, I. Watson. 1991. Simulation of surface urban heat islands under ‘ideal’ conditions in night. Part 2: Diagnosis of causation. Boundary-Layer Meteorol. 56: 339-358. Palmgren F. 1996. Design and Planning of Field Campaigns. In: Urban Air Pollution. Monitoring and Control Strategies.(I. Allegrini and F. De Santis. Eds.). pp. 77-99. NATO ASI Series/2, Environment, v.8. Ramsdell Jr. 1990. Diffusion in building wakes for ground-level releases. Atmospheric Environ. 24B: 377-388. Regulation No 8. National Standard for Ozone in air. Ministry of Environment. Ministry of Health. May 3, 1999 (State Journal No46-1999) (in Bulgarian). Roth M. 2000. Review of Atmospheric Turbulence over Cities. Quart. J. Roy. Met. Soc. 126: 941-990. Syrakov D. 1997. A PC-Oriented Multi-Level Eulerian Dispersion Model-Model Description. Bulg. J. of Meteorol. and Hydrology 8(1/2): 41-49. Syrakov E., K. Ganev. 1994. Transport, Chemical Conversion and Air Pollution Parameters on the Local Scales. Bulg. Geophys. J. 20(3): 5-25 (in Bulgarian). Takamura T. 1992. Spectral reflectance in an urban area. A case study for Tokyo. Boundary-Layer Meteorology 59: 67-82. Todhunter P., W. Terjung. 1988. Intercomparison of three urbam climate models. Boundary-Layer Meteorology 42: 181-205.
82
Ecology of the Sofia City CyanMagentaYellowBlack Page 83
Air Quality of Sofia City – Meteorological Aspects
Air Quality of Sofia City – Meteorological Aspects V. Andreev, D. Syrakov, A. Tsenkova, J. Ivantsheva, P. Videnov National Institute of Meteorology & Hydrology; 1784 Sofia, Bulgaria E-mail: [email protected]
SUMMARY The main factors and terms forming and defining air pollution are introduced. Urban air pollution management requires knowledge related to four important aspects: air quality monitoring network; meteorological data; emission inventories and numerical models. The main purposes, uses and requirements for the two kinds of air quality monitoring systems (an informative-warning and for general monitoring) are described and four dispersion model classes are examined. On the basis of mainly Bulgarian studies, results on the following questions on air pollution of Sofia are presented: the pollution potential of Sofia; lidar measurement results; an influence of the atmospheric parameters on surface air pollution; an air pollution distribution over the town territory; regionalization of Sofia according to air pollution levels; periodic fluctuations of air pollutant contents; space-time statistical structure of town air pollution fields; wind field deformations over complex relief areas; the influence of Sofia on air pollution of downwind areas; local scale air pollution modeling by the EPA Plume model, meteorological meso-scale and Eulerian dispersion models; a synoptical-statistical method for the air pollution level prognosis; and the chemical composition of precipitation in the town. KEY WORDS Air quality, monitoring network, meteorological data, emission inventories meteorological and dispersion models, lidar measurements, space-time statistical structure, synoptic-statistical prognosis method, Sofia, Bulgaria.
INTRODUCTION Air quality in a region is determined by two main factors - the sources of harmful emissions and the meteorological conditions at regional and local scales. These factors determine chemical transformations in the atmosphere. Air quality of the city of Sofia is worsened as a result of L. Penev, J. Niemelä, D. J. Kotze & N. Chipev (Eds.) 2004. Ecology of the City of Sofia. Species and Communities in an Urban Environment, pp. 83-127 © PENSOFT Publishers Sofia-Moscow
83
Ecology of the Sofia City CyanMagentaYellowBlack Page 84
V. Andreev et al.
mutual actions of these two major factors. One reason is the location of the city in a mountain valley. This is because meteorological conditions during a considerable part of the year are mainly characterized by calm weather and temperature inversions, which can be very intensive during winter. Another factor is the presence of many and various types and capacity sources of air pollutants since the city is a big residential and industrial centre. Contemporary knowledge of urban climate and air pollution are reviewed in the article “Urban climate and air quality – a review” by V. Andreev (this volume). The main findings are shortly represented here. Air pollution at an urban scale is the source of a range of problems: health risks mostly associated with inhalation of gases and particles, accelerated deterioration of building materials, damage to historical monuments and buildings, and damage to vegetation within and near the cities. In order to tackle these problems it is necessary to understand: what characterizes urban air quality; what are the sources of urban air pollutants; and which exposures and conditions are associated with the high concentrations of air pollutants occurring in cities. Air Quality Guidelines (AQGs) of the World Health Organization (WHO) are taken as reference values to assess where ambient concentrations may possibly affect human health and where further study may be necessary (see, e.g. Europe’s Environment, 1995). The significance of any pollution depends ultimately on the type of pollutants, the resulting exposure, health and other effects associated with this exposure. Pollutants, including noxious substances, in the urban environment may be “primary”, i.e. emitted directly from car traffic, domestic heating, industrial or other man-made and natural processes, but may also be “secondary”, i.e. the result of chemical reactions (photochemical and non-photochemical) in the atmosphere. Urban pollution results from a superposition from large scales, and from regional and local emissions from natural and man-made sources, transported and dispersed by meteorological processes at the same scales, and transformed by chemical reactions in the atmosphere. This is why the concentrations of air pollutants vary considerably with time (daily, weekly and seasonally) and in space. The actual occurrence, frequency and level of increased local urban air pollution concentrations depends primarily on: (a) the magnitude, business hours and spatial distribution of local emission sources situated close to the measurement/exposure location (e.g., traffic in the street) and (b) local climatic conditions (e.g., wind characteristics, frequency of calm weather, occurrence of inversion layers) that originate from the influence of regional climate on the local topography (e.g., flat terrain or basin or valley). Three main types of air pollution sources exist in urban areas: (a) large stationary sources with one or a few tall stacks (point sources), (b) small stationary sources distributed in an area (area sources) and (c) mobile sources, i.e. traffic (line sources). Only the fastest chemical reactions are important for locally emitted pollutants. Their concentrations are mainly determined by dispersion. A specific feature of air pollution, especially in cities, is the existence of short-lasting and comparatively high concentrations of pollutants, called “episodes”. These may last from one to several days, and should be discerned from “momentary” peaks of several hours’ duration, and from the “background” variability of monthly and seasonal character. Episodes are associated with various meteorological factors such as poor ventilation or, in some cases, unfavorable transport and transformation processes. There are two kinds of episodes with 84
Ecology of the Sofia City CyanMagentaYellowBlack Page 85
Air Quality of Sofia City – Meteorological Aspects
high concentrations of pollutants – “winter episodes” (“sulphuric acid smog episodes”), characterised mainly by an increase of SO2, and suspended particulate matter (SPM) concentrations, and “summer episodes” (“photochemical smog episodes”), characterised by the production of ozone – O3. Winter episodes occur when low winds and a strong temperature inversion impede the vertical mixing and dilution of pollutants in the lowest atmospheric layers. During these episodes, emissions related to domestic heating can be up to 70 % higher than the winter season average. As well as SO2 and SPM, the winter smog mixture contains compounds such as carbon monoxide (CO), nitric acid (HNO3), nitrogen dioxide (NO2) and aerosol particles of variable content such as soot, sulphate, nitrate, ammonium, metals and organic compounds, such as polycyclic aromatic hydrocarbons (PAH). Under winter smog conditions, daily average SO2 concentrations in rural areas can reach 400 µg/m3; it may be higher in cities (e.g. in December 1952, daily average concentrations of SO2 and SPM in London reached values of about 5000 µg/m3 each). It is essential to note that the winter smog episodes often extend across several European countries due to long-range transport across the continent. Photochemical smog episodes occur under conditions of high insolation and low winds, when the concentrations of ozone and photochemical oxidants can build up to high levels which are damaging both to human health and to vegetation. The main cause of photochemical oxidant and ozone formation are the emissions of NOx (NO+NO2) and VOCs (Volatile Organic Compounds) from motor vehicles, which are densely concentrated in city areas as well as from industry (in particular, combustion of fossil fuels, solvent use and evaporation losses). VOCs present in the air from anthropogenic and natural emissions are oxidized in the atmosphere in the presence of sunlight. The process generally starts through reaction with the very reactive hydroxyl (OH) radical, by reaction with ozone or by direct action of sunlight. The oxidation processes are complex, and NOx play an important role as necessary catalysts. A host of secondary products is formed during this oxidation; they are usually called photo-oxidants. Ozone is by far the most important of these in terms of the adverse effects it has on human health and ecosystems. Several international bodies such as the WHO, UNECE (UN Economic Commission for Europe) and the EC (European Community) proposed or set guideline values for ozone for the protection of human health and/or vegetation. These guideline values indicate levels combined with exposure times during which no adverse effects are currently expected. For example, the 150 µg/m3 hourly concentration level and 100 µg/m3 for 8 hours are proposed for the protection of human health (for ozone, 1ppb » 2µg/m3). Bulgarian guideline values for ozone are: 110 µg/m3 – (average value for 8 hours)- for human health protection and 200 µg/m3 (hourly averaged) – for the protection of vegetation. The hourly averaged values of 180 and 360 µg/m3 are the guideline for information or alert for the inhabitants, respectively (see, Regulation No8, 1999). Progress in scientific understanding may lead to revision of these guideline values. The photochemical smog episodes usually take several hours to develop. But such situations are not only local scale phenomena. Because of the strong link to particular meteorological situations, the occurrence of “ozone episodes” in Europe is very variable from year to year in both space and time. Oxidation takes place on time-scales ranging from a few hours to some85
Ecology of the Sofia City CyanMagentaYellowBlack Page 86
V. Andreev et al.
times, several months. During this period, the polluted air is transfered due to long-range transport. That is why the influence of these episodes on air quality and deposition of pollutants can extend over hundreds or several thousands of kilometers, i.e. they are as well regional or continental scale phenomena. The term “hot spot” pollution may be used to describe the high short-time pollution concentrations to which the population may be exposed when located close to pollution sources. “Hot spot” pollution includes the case of urban streets with busy car traffic, and the pollution impact from industrial stacks in cities. In all European cities, smog occurrences and long-term average concentrations of harmful compounds such as lead, benzene, SPM and benzo(a)pyrene (BaP) receive a significant contribution from emissions from road transport. Benzene and BaP are indicators of organic compounds associated with volatile organic hydrocarbons and with soot and PAHs from small-scale combustion processes, respectively. Road transport contributes, on average, more than half to the nitrogen oxides and about 35% of the VOC emissions in Europe. While NOx emissions from stationary sources have decreased in many urban areas, emissions from motor vehicles have increased because the growth in the number of vehicles and the distance traveled per vehicle has been much larger than the reduction in emission factors. Since the lifetime of vehicles is typically more than 10 years and in some countries such as Bulgaria more than 20 years, reductions of emissions will be slow even if vehicle numbers and distances traveled were to stop increasing. More rapid improvements in air quality in the short term can only be achieved by limitations on road transport in urban areas. In many cities, air pollution from road transport and domestic heating is supplemented by that from local industry. The impacted area is usually a few kilometers from the emission sources, depending on the height of the stacks and the prevailing wind directions. Monitoring stations in most cities are operated in areas where principal industries are expected to have a large impact. These “hot spot” stations indicate a significant impact from industrial emissions, i.e., air concentrations of SO2 and/or PM are considerably higher than in the city center. Such distribution of pollutant concentrations should often be observed in some industrial zones of Sofia. Most of the primary pollutants from cars, domestic heating and some industrial sources are emitted and transformed by the chemical reactions at ground level where severe pollution levels might be found. However, people live and realize the prevailing part of their own activity in the surface layer of the atmosphere. That is why, for the preservation of a favorable air quality in urban areas, anthropogenic emissions must be controlled. Successful measures for reducing air pollution in cities over the past 20 to 30 years have included the regulation of fuel for domestic heating (low sulphur content in oil and coal, introduction of natural gas instead of coal and oil), development of district heating, reduced lead content in petrol and restriction on the use of cars in the city centers. Furthermore, dust removal from industrial sources and large boilers have resulted in major improvements. These measures have been particularly successful in reducing SO2, particle and lead emissions, and hence daily and annual average urban atmospheric concentrations, by significant amounts in many cities. However, there has been little or no evidence of similar downward trends in NOx concentrations. During smog conditions, emergency actions are taken by some cities (e.g., Rome, Milan, Athens, Prague) involving, for example, restrictions on car traffic and fuel substitution in power plants/ 86
Ecology of the Sofia City CyanMagentaYellowBlack Page 87
Air Quality of Sofia City – Meteorological Aspects
industries. Cost-effective energy conservation measures and substitution of fossil fuels with renewable energy resources have a large potential for the reduction of emissions. As a result of such simple premises, urban air pollution management requires knowledge of the parameters related to four important aspects: a) air quality monitoring network, which provides measurements and data relevant to several pollutants, their occurrence in time, their spatial distribution and, eventually, episodes exceeding the quality standards; b) meteorological data not only from the surface layer but also from the boundary layer in the meteorological network; c) emission inventories, which provide information on the emitting source and their temporal evolution; d) numerical dispersion models, which provide information on the prediction of air quality from emission and meteorological data. Air quality and the state of the atmosphere are usually fixed through corresponding networks for air quality monitoring and for meteorological measurements. They are composed of many measuring stations. One of the main purposes of these networks is to produce information for understanding the basic processes, which are responsible for atmospheric pollution and to take measures required for reducing the risk for health and the environment from air pollution. Two kinds of permanent networks and systems for air quality monitoring are used in dependence on the necessity to take either long-term or short-term measures for reducing the risk from pollution. They are: networks (systems) for general monitoring of air quality, and air pollution informative-warning systems (networks). General monitoring systems are considered the basic systems in urban areas. Data produced in the system are used for long-term environmental programs; for assessment of the effectiveness of a long-term policy, of the air pollution level in connection with the national ambient air quality standards, of the trends and validation of mathematical models; for documentation of population exposure to air pollution; for town and country planning; for scientific investigations etc. The stations of the urban general monitoring systems are located in inhabited areas, where the expected long-term concentrations of a particular pollutant are higher than in other areas of the town as well as where concentrations of a particular pollutant may be treated as average for densely populated parts of the town. These kinds of stations can also provide feedback for many other specific purposes. The informative-warning systems are of two kinds – warning and alarm systems. Warning systems are based on air-pollution forecast, which, in most cases are for information only. Alarm systems are based on actual measured concentration values, which are connected with enforcement of compulsory preventive actions. These systems collect data in real time by means of continuous monitors. The main task of them is to provide data for short-term pollution abatement procedures. They should be established only in densely populated areas with frequent occurrence of high concentrations of pollutants caused by different sources and especially from a particular industrial plant or from distant plants as well as from traffic. The meteorological data, necessary for the functioning of the informative-warning systems are received through meteorological observations either at the same point where air pollution is measured, or at the nearby meteorological station. During field experiments these two types of data - for the concentrations of pollutants and atmosphere parameters, are most frequently conducted simultaneously at one and the same points. It can be noted that during episodes, extraordinary measures can be applied to protect the population. They are two kinds: active measures – emergency reduction of local emissions, and passive measures for a reduction of health hazards. 87
Ecology of the Sofia City CyanMagentaYellowBlack Page 88
V. Andreev et al.
The procedures for the realization of emission inventories are well established. It can be noted that there is no sufficiently comprehensive and well-founded emission inventory for Sofia city. The processes that determine the distribution of gas pollutants in air and that are considered in the atmospheric dispersion models are: advection and convection (horizontal and vertical transport); diffusion into the three directions in space; dry deposition at the moment of contact of pollutants with the earth surface; the washing off of pollutants by precipitation (wet deposition); chemical transformations and/or radioactive decay. When having pollutants, carried by aerosol particles one must additionally include the processes of gravitational sedimentation and transformation of particles. These depend on the spectrum of particle size that itself changes in time as a result of interactions among particles in a gas or liquid environment (processes of decay or coagulation). These processes, as well as the chemical transformations, develop differently in the dry atmosphere, in a cloud environment and in precipitation. Adding the fact that as a result of chemical transformations, some gasses transform into aerosol particles and vice versa, we can obtain a notion of the great complexity of the distribution of air pollutants. Each of the latter processes can be described mathematically with a different degree of complexity and with different time and space characteristics (scale and resolution). It is obvious that with such a number of processes in a continuous medium like the atmosphere, the only way to describe them is through numerical modelling. In the international literature, a great number of numerical dispersion models have been created with different purposes and degrees of complexity at realization. The different models can describe the distribution of one or many pollutants emitted in the atmosphere by different types of sources at various time and space scales. In accordance with the approach for the description of dynamic processes, the dispersion models can be separated generally into 4 classes: Gaussian plume models; Lagrangian-Eulerian cloud models (puff-models); Eulerian models; Lagrangian particle models. The nature and main characteristics of each of these types of models are described in the article “Urban climate and air quality – a review” by V. Andreev (this volume). Urban ground is covered with natural and artificial materials, buildings and other obstacles. The heat budget as well as the flow structure close to the ground depend largely on the height and density of the constructions. Flow structure varies from a series of turbulent wakes due to isolated obstacles to strongly interacting turbulent streaks and even to turbulent flows having some similarities with flows in semi-porous media. The atmospheric layer below roof level is called the urban sub-layer or urban canopy by analogy with the atmospheric layer between ground and the trees of dense forest. The dispersion of pollutants in the canopy layer has a primary importance because human life and activity are mainly in this sub-layer. But experimental data obtained inside large cities show that the vertical structure of the urban low atmosphere as a rule is very complex. Its multiple structure largely dependent on the structure of the canopy elements (see Andreev, this volume). One may suppose, that new on-site experiments with a large concentration of sensors allowing extremely detailed documentation of flows and climate in typical canopy geometries are necessary to validate the available mathematical models. However, according Mestayer and Sandrine (1995), it appears that the multiplication of measurements will not suffice to assess and to predict the flow structure, the dispersion processes and the air quality or pollutant contents, and 88
Ecology of the Sofia City CyanMagentaYellowBlack Page 89
Air Quality of Sofia City – Meteorological Aspects
their impacts because of the large structural, temporal and spatial variability of the urban lower atmosphere. The detailed prediction and real time control of air quality call for the development of a new generation of complex numerical models combining physics and photochemistry of the urban atmosphere. This will aid in the understanding of the relations between the physical-geographical conditions, the regional meteorology, the local pollutant sources and the monitoring measurements. The development of sub-meso models will allow for the study of the interactions of the city fabrics and urban sources of heat and pollutants with the surrounding orography. This is of special importance for cities like Sofia, located in valleys or valley systems, where the urban radiative budget interacts with the along-valley and katabatic cross-valley winds; this can produce extreme pollutant traps. This model development is possible only if the structure of the flow in the intermediate layer just above the roofs is much better understood. Urban planners, local decision-makers and city leaders, as well as responsible officials of the environment, need more complicated data to understand the results of their policy, especially concerning traffic management, to assess the actual air quality with regard to secondary pollutants, to finely predict all-type impacts of new chemical sources, as well as to foresee the effects of new regulations. The need for estimations of secondary pollutant concentrations is especially demanding, since they depend altogether on mixing and reactive processes: this implies the simultaneous knowledge of chemical kinetics and of transport-dispersion behaviour of primary and intermediate compounds.
MATERIAL AND METHODS Present knowledge of air pollutant dispersion is based on three types of data: on site measurements; numerical model results, and wind tunnel and water tank simulations. It must be noted that a significant part of this knowledge also comes from field campaigns and even from incident results of research and observations in the neighbouring fields of dynamic meteorology, air flow over complex terrains and wind engineering. We must add the useful measurements of the urban atmospheric phenomena by lidars, sodars, radar profilers and satellites. Perennial data from the National network for air quality monitoring should be used to characterise the concentration fields of air pollutants in the Sofia hollow valley and in Sofia City. Systematic measurements of the air quality in Sofia City and around the metallurgical plant “Kremikovtzi” have been conducted since 1969, first by the Hygiene and Epidemiology Inspection and the Hygiene Center, and later by NIMH (since 1973) and the Ministry for the Environment and Waters. The number of stations in the monitoring system in Sofia changed, but reduced substantially during the recent 10-15 years. This limits the variety of scientific research in this field. Data from the general monitoring network have many methodical peculiarities, marked in official state documents (see “National Action Plan on Environment and Health...”, 1996 or “National Environmental Strategy and Action Plan 20002006”, 2001). Some of these, in connection mainly with the duration and frequency of manual air sampling, are inadequate when it is necessary to use the data for other purposes, as meteorological ones for example. The most important peculiarity in the functioning of the 89
Ecology of the Sofia City CyanMagentaYellowBlack Page 90
V. Andreev et al.
stations with manual manipulations, is that air sampling continues to be performed only during working days with a frequency of four times during daytime (see Regulation No 8, 1999). This means that all mean concentration values determined on these measurements are incorrect - the mean diurnal, because there are no measurements at night; the mean monthly (or weekly), because no measurements are conducted during weekends and holidays. Naturally, all other mean values for longer periods (seasonal, yearly) cannot be properly calculated with this measurement regime. Therefore we do not have at our disposal representative perennial mean air pollutant concentration values from a 24 hours or longer period. It is clear that with this regime, no representative values can also be obtained for extreme concentrations (the maximal ones) from single samplings. The use of data from the monitoring network was considerably limited with the introduced by Regulations No 8 and No 9 (1999) (one-hour duration) of the single manual samples for SO 2, NO2, ozone and fine dust particles (the time for averaging for the automatic stations is also one hour). It is, probably, suitable for the purposes of healthcare, but it is not suitable for other purposes like meteorological. The analysis of turbulent energy fluctuations of the meteorological elements shows that the most stable mean values for air characteristics are obtained at an interval of between 10 and 30 minutes. This is why the introduced one-hour sampling, instead of the former 30minutes, leads to an uncertainty in comparing the measured concentrations with the models. The general settings show (see Palmgren, 1996 or Walczewski, 1996) that when a general monitoring network for air quality is being constructed in an urban or an industrial center, attention is paid mainly to finding the level and peculiarities of air pollution in the whole city. When a study of this problem is necessary for a local area in the city, this network can be expanded with an increased density for that area. This was done, for example, by the sanitary agency (HEI)-Sofia in the region of MP “Kremikovtzi”. This approach, however, is expensive and is rarely applied. The air quality in local areas is studied most frequently by “field (natural) experiments”, also called “Field campaigns”. They are necessary during the initial stage of creating general monitoring networks or in reorganization or regular optimization of functioning networks. The experiments are usually conducted for representative periods (from 15 days to one or a few months) from the cold and warm half-year periods by increasing station density and frequency of measurements, accompanied obligatorily by ground and higher level measurements (Palmgren, 1996). Such experiments are necessary in research on air pollution in Bulgaria at least because of the marked peculiarities of the data from the national monitoring network. In this publication the measurements on air pollution in Sofia City will be reported that are obtained on the basis of short-term experiments encompassing the whole city and are noted as “Sofia’76” (Andreev et al., 1982) and “Sofia’82” (Andreev, 1984) or parts of it (SofiaDrujba’93) (see Syrakov, 1992-1994). The experiments were conducted during typical winter (25.11.-9.12.1976; 20.11-20.12.1981 and 15-26.11.1993) and summer (10-25.06.1976; 15.0615.07.1982 and 10-30.06.1992) periods respectively at 14 and 19 points in the whole city or 8 points in the Iskar district (Drujba). The concentrations of air pollutants in Bulgaria are measured only at ground level. The measurements at altitude are realized only with lidar systems by methods developed in the Institute of Electronics of BAS (see Donev et al., 1995; Kolev et al., 2000; Syrakov D., 19921994 - report for I stage). The use of lidar systems in a range of cities in Bulgaria, including 90
Ecology of the Sofia City CyanMagentaYellowBlack Page 91
Air Quality of Sofia City – Meteorological Aspects
Sofia, shows that at least at present this is the best method for the diagnosis of the distribution of layers of aerosol particles in the air, as well as its change in time. This allows for the documentation of intensive emissions by local sources and to register smoke plumes. In our opinion it is important to measure the absolute concentrations of aerosol particles in air using lidar systems. Such measurements were conducted, for instance, in 1992 in the Iskar district, and in the report for the I stage (Syrakov, 1992-1994) the horizontal and vertical sections of the aerosol concentration fields at different altitudes during different time periods are presented. A method should be found for comparing such results with data by direct measurements of the concentrations. In this sense, the use of a “relative” concentration, as it was done by another group of researchers also from the Institute of Electronics during measurements in 1994 and later (see Donev et al., 1995; Kolev et al., 2000), is better. In the National Institute of Meteorology and Hydrology (NIMH) some versions of most of the basic classes of mathematical models for the distribution of pollutants in the atmosphere were developed. Thus, for example, a set of Gaussian models, built on formulas and parameterizations recommended by the Environment Protection Agency (EPA) of the USA is available. The most important characteristics of these models are as follows: a) Stability is determined depending on the meteorological conditions in accordance with the classification suggested by Pasquil; b) formulas of Briggs are used for the dispersion functions and for the elevation of the plume over the opening of the stack in result of inertial and Archimedean forces; c) wind speed at the stack opening is determined by the ground wind and the power law recommended by the EPA; d) in the dispersion functions, a correction of sampling time is introduced at concentration measurements, taking into account the effect of the plume meandering; e) the model takes into account the altitude of the mixing layer if such data are available; f) calm weather conditions are treated as the plume formed at wind speeds of 0,5 m/s in all directions. In the case of a gas pollutant, the processes of dry and wet deposition are taken into account, as well as the decrease of the pollutant’s mass as a result of chemical transformations. Modules for calculating the distribution of aerosols are introduced – gravitational deposition and specific washing. As one of the pollutant’s characteristics, along with the speed of dry deposition, the mean speed of degradation (radioactive decay) is also introduced. The transformation processes that the pollutants experience are also calculated in the models. The dry and wet deposition not only lead to a decrease of the pollutants’ concentrations in the air, but also form a field of accumulation on the earth surface. Entry into the models is interactive, but using the capacity of the Disk Operational System, the responses to all the questions put by the program can be saved in advance in a text file that can be specified as the entry file. The meteorological entry is both a situation series and climatic information. The result of the models is a GRD file of the SURFER graphics package in text format. Thus it can be used both as numerical values and the field can be presented graphically as an isoline or a 3D surface. NIMH has a well developed puff-model, called LED (Lagrangean-Eulerian Diffusion) that is used for solving different problems such as pollution with sulphur and nitrogen oxides, formation of acid rains, continental and regional pollution with heavy metals, etc. In the LED model the horizontal distribution of the concentration is Gaussian, while the vertical distribution is different depending on the boundary conditions on the earth surface and at the upper 91
Ecology of the Sofia City CyanMagentaYellowBlack Page 92
V. Andreev et al.
boundary of the mixing layer (if it is present). In the model there are calculated the other processes accompanying the transport and diffusion - the dry deposition, the washing by precipitation and the chemical or radioactive transformations. In order to provide data for wind in the center of every puff, and also data for the diffusion capacity of the atmosphere, a relatively simple, but effective model of the atmosphere boundary layer is included, that is supplied with standard meteorological data from only two levels - the earth surface and the upper boundary of the boundary layer (approximated with a level of 850 hPa). For feeding the LED dispersion model, meteorological data obtained by the global prognostic Brecknell model is used in a network of 2.5 at 2.5 degrees. In NIMH an Eulerian model is available, called EMAP, an abbreviation from Eulerian Model for Air Pollution. Up to now, a few versions of EMAP have been proposed – for describing the distribution of an inert gas, of sulphur compounds (SO2, SO4=), of heavy metals and long-life air pollutants carried by aerosols (Pb, Cd, Benz(a)pyrene, etc.), and of mercury. EMAP is a threedimensional model for numerical simulation of a range of processes determining the distribution of air pollutants. Here a Cartesian coordinates system is used with an invariable step on the horizontal and variable step on the vertical. The horizontal coordinate surfaces that follow the relief shape are denser at ground level and are rarefied in altitude corresponding to the typical profiles of the meteorological elements in the atmospheric boundary layer - quick changes near the ground and smoother changes in altitude. A detailed presentation of this model is presented in BC-EMEP (1995) and Syrakov (1997). The EMAP model is used in important tasks like representing Bulgaria in the European EMEP program or the NIMH system for early warning in the case of a major nuclear accident that supplies Civil Defense and the Crisis Team of the Bulgarian Committee for Nuclear Energy Utilization with a 3-days prognosis for the evolution of the radioactive pollution patches. Until 2-3 years ago, a special modification of the LED puff-model was used for this purpose. The great variety of dispersion models poses questions like which type of models to choose from and to what degree the results from the different models could be trusted. The answer to the first question can be given only by experts in that field and depends on a range of factors - the character of the task, the available equipment, information base and models. The second question is the subject of intensive experimental, theoretical and modeling activity at an international scale. The process of comparing the modeling results with experimental ones is called verification, validation or inter-calibration. Without discussing the differences among the three terms we must note that considerable effort was put into creating the methods for such comparison, showing that it is not an elementary task. The problem is that, on one hand, the measurements that are usually conducted at one point in space are burdened with inevitable measurement errors. These errors are even greater for field experiments where measurements are done simultaneously at a large number of points with simpler equipment than can be installed in laboratory conditions. On the other hand, the modeling results are usually not related to a fixed point, but characterize some area around that point having sizes equal in resolution. All these considerations must be regarded in comparing results.
92
Ecology of the Sofia City CyanMagentaYellowBlack Page 93
Air Quality of Sofia City – Meteorological Aspects
RESULTS Main air pollution sources in Sofia City. According to the Concept of the social-economic development of Sofia up to 2020 (see General Structural Plan of Sofia, March 2002) about 27% of the territory of Sofia city are production areas, while only 5-8% in the capitals of advanced European countries. It is a characteristic for Sofia City that 34% of the production areas are located no more than 4 km from the city center and a considerable number of enterprises are within residential districts or in close proximity to them. According to this Concept, the production zones in Sofia are combined in the following 3 large industrial complexes: (a) Northern industrial complex including the production zones Levski-north, Hadji Dimiter-Malashevtsi, Area rear the Central railway station, Voenna rampa-west and east, Iliantsi-west and east, Novi Iskar-Kurilo, Svetovrachane; (b) Eastern industrial complex from the production zones Iskar-north and south, Kazichene, Hristo Botev-Slatina, Gorubliane, Izgrev, Vitosha, Sredets; (c) Western industrial complex from production zones Zaharna fabrika, Orion, Lulin, Voluiak. Similarly, around the city there are 3 large industrial and transport complexes namely Kremikovtsi-Gara Iana, Gara Elin Pelin-Ravno pole, along a direction Sofia-Slivnitsa. The contribution of transport should be noted, because a great part of the country’s automobile park is used in Sofia. This park consists mainly of old cars (mean age 14 years, 8-10 years for busses). They travel on roads having technical characteristics (dimensions, coverings, radius, length slope etc.) that do not comply with normative requirements and are badly maintained. The result is a decreased traffic capacity, traffic jams and accidents. That is why the auto-park in Bulgaria and Sofia in particular has a low ecological index, i.e. it is one of the most serious air pollutants. Another emission source is domestic heating. It is characterised by inefficient heater that are mainly concentrated in the districts with individual houses located on the northen slopes of the Vitosha mountains as well as in Lozen, Malinova dolina, Vitosha, Krastova vada, Manastirski livadi, Gorubliane, Orlandovtsi, Malashevtsi, Fakulteta, Hristo Botev and Iliantsi. Furthermore, dust, in considerable quantities, is present in the air because of the bad status of streets during times of communal and repair activities. These various sources, concentrated in a relatively small area, cause a comparative high level of the air pollution in Sofia city. The concentrations of air pollutants can exceed the sanitary standards during adverse meteorological conditions. Air pollution levels in a certain area is determined by two major factors – the parameters of the sources of harmful emissions and the meteorological conditions at regional and local scales. Air pollution, formed under the influence of the climate and microclimate of an area, assuming that the sources’ parameters remain constant, can be defined as the “air pollution potential”. The latest assessment of this potential in Sofia City was conducted by M. Teneva, P. Ivanov and others in 1980’s (see Blaskova et al., 1983; Teneva & Batchvarova, 1989), applying a physical-statistical approach created at the General Geophysics Observatory–St. Peterburg, Russia. Data for the frequency of ground temperature inversions were used along with the frequency of weak winds (up to 1m/s) and fog, as well as from a of combination of these factors. According to this approach, the relative pollution potential during city conditions, calculated for low sources and cold emissions, is classified into 6 degrees (Table 1). In Table 2 93
Ecology of the Sofia City CyanMagentaYellowBlack Page 94
V. Andreev et al. Table 1. Degrees of the Air Pollution Potential (APP). Degrees
Level of APP
Interval of APP
I II III IV VA VB
Low Moderate Increase High Very high Unusually high
0,0÷2,4 2,5÷2,7 2,8÷3,0 3,1÷3,3 3,4÷3,6 3,7÷4,0
Table 2. Air Pollution Potential of Sofia. Hour
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
XII
Mean
02h 08h 14h Mean
3,7 3,8 3,1 3,6
3,4 3,7 2,6 3,0
3,3 3,1 2,4 2,8
3,2 2,8 2,2 2,7
3,5 2,8 2,4 2,8
3,7 2,8 2,4 2,8
3,7 2,8 2,4 2,8
3,6 2,9 2,4 2,8
3,9 3,0 2,4 3,0
3,7 3,5 2,5 3,0
3,8 3,8 2,9 3,4
4,0 3,6 3,1 3,7
3,6 3,1 2,4 3,1
mean values are presented for three climatic terms of observation (02, 08 and 14h local time) and for the whole 24 hours period at station NIMH-Sofia. In Sofia City at 02h at night, during the whole year, climatic conditions are present that help the retaining of compounds in the surface air. Due to this, the pollution potential throughout the year appears very high (V A degree), and except for the February-May period it is exceptionally high - VB degree. The potential in Sofia at 08h in the morning is very high for the April-September period, and during the cold half-year period - high and very high. The conditions for dispersion of compounds are considerably better at 14h and during the greater part of the year the potential is moderate; only for the NovemberJanuary period it is high. Discussion on air quality in a certain region begin with characterizing the concentration fields of the main air pollutants. It is reasonable to use generalized indicators such as, for example, mean concentrations defined by perennial observations in a network of stationary air quality monitoring points. As it was already pointed out, this cannot be done for any cit